Advanced Science - 2020 - Hettich - Exosomes For Wound Healing Purification Optimization and Identification of Bioactive

FULL PAPER
www.advancedscience.com
Exosomes for Wound Healing: Purification Optimization

and Identification of Bioactive Components
Britta F. Hettich, Maya Ben-Yehuda Greenwald, Sabine Werner,
and Jean-Christophe Leroux*
Exosomes are a therapeutically relevant

Human mesenchymal stem cell exosomes have been shown to promote EV subtype given their activity in a broad
cutaneous wound healing. Their bioactivity is most often attributed to their range of disease categories, including
protein and nucleic acid components, while the function of exosomal lipids immune-related disorders and tissue
remains comparatively unexplored. This work specifically assesses the regeneration.[3,4] They correspond to a
heterogeneous population of naturally
involvement of lipids and the transmembrane enzyme CD73 in the exosomes’
occurring, endosome-derived nanovesicles
biological activity in stimulating the cutaneous wound healing process. Since with a maximal diameter of 200 nm, and
exosome preparation processes are not harmonized yet, certain production are constituted by a protein-containing
and purification parameters are first systematically investigated, enabling the lipid bilayer, which encapsulates nu-
optimization of a standardized protocol delivering high exosome integrity, cleic acids, proteins, and other bioactive
components.[1,2,5,6] Exosomes can be viewed
yield, and purity. An in situ enzymatic assay is introduced to specifically
as cellular fingerprints, as their composi-
assess the vesicle functionality, and quantitative proteomics is employed to tion is highly dependent on the producing
establish the cell culture conditions yielding a stable exosome protein profile. parent cell, and dictates the biological
Using a combination of in vitro approaches, CD73 and constitutional lipids of processes these natural vesicles can trigger
HPV-16 E6/E7 transformed human bone marrow stromal cell-derived at target sites.[1,2,7] In contrast, certain
exosomes are identified as key bioactive components promoting the proteins’ absence or presence can be used
as markers of exosome identity and purity,
exosome-driven acceleration of processes required for wound repair. A pilot
independent of the cell source.[1,2]
wound healing study in mice indeed suggests a role of lipids in the exosomes’ Exosomes derived from human mes-
biological activity. Strikingly, the extent of the bioactivity of these exosomal enchymal stem cells (hMSC) have been
components is found to be dependent on the target cell type. sporadically reported to improve cutaneous
wound healing through the modulation of
inflammation, proliferation and/or matrix
remodeling.[3,4,8] Certain proteins and microRNAs carried by the
1. Introduction exosomes have so far been proposed as key mediators of these
effects.[9–11] The transmembrane enzyme ecto-5′-nucleotidase
Cell-secreted extracellular vesicles (EV) have been identified (CD73) expressed by hMSC exosomes has been associated to
as key players in intercellular communication processes.[1,2] their beneficial immune-modulatory function in graft-versus-
host-disease[12] and capability to promote cartilage repair,[13] also
suggesting a potential role for CD73 in the hMSC exosome-
B. F. Hettich, Prof. J.-C. Leroux
Institute of Pharmaceutical Sciences driven acceleration of skin wound healing. Furthermore, a pre-
Department of Chemistry and Applied Biosciences vious study identified exosomal lipids as key components of
ETH Zurich the exosome effects in anticancer therapy.[14] Indeed, certain
Zurich 8093, Switzerland lipid classes (e.g., sphingolipids, glycerophospholipids) can ex-
E-mail: [email protected]
ert immune-modulatory, mitogenic, migratory, or angiogenic ef-
Dr. M. Ben-Yehuda Greenwald, Prof. S. Werner
Institute of Molecular Health Sciences fects, which have been proven beneficial in impaired cutaneous
Department of Biology wound healing.[15–17] The function of specific lipids in mediating
ETH Zurich the exosomal therapeutic efficacy was, however, hardly investi-
Zurich 8093, Switzerland gated in past studies.[18]
The ORCID identification number(s) for the author(s) of this article
Currently, progress in exosome research is in part hampered
can be found under https://doi.org/10.1002/advs.202002596 by the heterogeneity in the applied preparation procedure.[19,20]
© 2020 The Authors. Published by Wiley-VCH GmbH. This is an open The significant variations in the purification efficiencies of the
access article under the terms of the Creative Commons Attribution different isolation methods (i.e., the enrichment of exosomes
License, which permits use, distribution and reproduction in any compared to other EV and impurities) prevents, in most cases,
medium, provided the original work is properly cited. the generalization of findings across different studies.[20–22] To
DOI: 10.1002/advs.202002596 date, an array of different state-of-the-art exosome preparation
Adv. Sci. 2020, 7, 2002596 2002596 (1 of 15) © 2020 The Authors. Published by Wiley-VCH GmbH
21983844, 2020, 23, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/advs.202002596 by Nat Prov Indonesia, Wiley Online Library on [02/09/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.advancedscience.com
Figure 1. UC purification yields a higher exosome recovery and purity than UF-SEC. A–C) HS-5 exosomes isolated by either A) UC at 150 000 × g or
B) by UF-SEC after a 24-h production period. A,B) NTA size profiles (left panels) and TEM images (right panels) of exosomes. Green arrows indicate
the main vesicle population with their modal diameter and red arrows indicate vesicle aggregates (panels (A) and (B)). C) Western blot analysis of the
exosome marker proteins TSG101, CD9, CD63, and CD73, as well as the contamination marker calregulin, with GAPDH as loading control. Experiments
(left panels in (A) and (B); panel (C)) were performed in triplicates and a representative plot/image is shown.
strategies is available and a growing number of publications has the recommendations of the International Society for Extracellu-
investigated their assets and drawbacks.[6,23,24] Ultracentrifuga- lar Vesicles.[30,31]
tion (UC) was the first exosome isolation method to be estab- With 5.4 ± 1.8 × 1010 vesicles mL−1 , UC resulted in an ap-
lished and although it remains the gold standard,[25] drawbacks proximately twofold higher vesicle recovery than UF-SEC, which
such as copurification of protein aggregates, morphological yielded 2.9 ± 1.5 × 1010 vesicles mL−1 . As measured by nanoparti-
changes (e.g., agglomeration, membrane rupture), and tedious cle tracking analysis (NTA), the UC-isolated vesicles had a modal
processing times have motivated the implementation of faster diameter of 139 ± 6 nm, while those obtained by UF-SEC were
alternatives.[6,24] Ultrafiltration (UF) is a more rapid purifica- smaller, measuring 121 ± 6 nm (Figure 1A,B). Further vesi-
tion method with a proposedly higher purification efficiency,[23,26] cle/aggregate populations ranging up to 700 nm were detected
which when in combination with gravitational size exclusion with both purification approaches, albeit in low proportions.
chromatography (SEC) can produce high yields of pure and in- The size distributions of UC- and UF-SEC-purified exosomes
tact exosomes.[27–29] obtained by NTA were confirmed by transmission electron mi-
In this work, the biological activity of HPV-16 E6/E7 trans- croscopy (TEM), which also revealed the presence of aggregates
formed human bone marrow mesenchymal stromal cell (HS-5)- in both cases (Figure 1A,B; Figure S1A,B, Supporting Informa-
derived exosomes on different aspects of cutaneous wound heal- tion). Moreover, the vesicles harvested by either isolation method
ing was investigated, focusing on the role of previously largely exhibited a distinct cup-shaped morphology, which has been pre-
unexplored exosomal components, inter alia CD73 and exosomal viously reported as an artefact of the preparatory fixation process
lipids, in mediating the exosomes’ regenerative capability. To this required for TEM.[23]
end, a standardized purification protocol was first optimized to Immunoblotting revealed that both isolates were enriched in
reproducibly obtain intact and bioactive HS-5 exosomes in high the exosome marker proteins TSG101, CD9, CD63, and CD73
purity and yield. (Figure 1C). The UC preparation was devoid of the endoplasmic
reticulum marker protein calregulin, a representative contami-
nant, whereas the UF-SEC sample was not. Higher impurity lev-
2. Results els following UF-SEC were also reflected in a higher protein con-
tent (161 ± 97 µg protein mL−1 for UF-SEC vs 95 ± 20 µg protein
2.1. Exosome Purification by UC Outperforms UF-SEC mL−1 for UC) and a reduced vesicle-to-protein ratio (1.9 ± 0.3 ×
108 vesicles µg−1 for UF-SEC vs 5.6 ± 1.2 × 108 vesicles µg−1 for
The purity and functionality of exosomal formulations is highly UC). The latter has been previously proposed as a measure of the
dependent on the exosome isolation method.[20–22] In order to ob- vesicle purity.[26,32]
tain a high yield of enzymatically active exosomes, which can Moreover, the CD73 enzymatic activity following UC isolation
be reproducibly used for in vitro and in vivo wound healing was tenfold higher than that after UF-SEC (4.4 ± 0.7 U mg−1 for
studies, the HPV-16 E6/E7 transformed human bone marrow UC vs 0.4 ± 0.1 U mg−1 for UF-SEC, whereby 1 unit (U) is defined
mesenchymal stromal cell line HS-5 was selected for a detailed as 1 µmol min−1 ). This could indicate a partial activity loss upon
investigation of two common purification strategies. Exosomes the UF-SEC procedure or, in line with Western blot analysis, a
were produced by HS-5 cells for 24 h under serum-free condi- lower purity of the UF-SEC sample (i.e., less CD73 per isolated
tions in order to obviate a possible contamination with serum- µg of protein).
derived components such as EV, lipids, and RNA.[1,23] Vesicles Taken together, the vesicle characteristics (i.e., size, mor-
were subsequently isolated by either UC at 150 000 × g or UF- phology and presence of exosome protein markers) were
SEC and the isolated fractions were characterized according to well inside specifications for exosomes,[30,31] demonstrating the
Figure 2. The preparation conditions impact the exosome yield and purity. A) Scheme illustrating the exosome preparation process. Investigated pro-
duction and purification parameters are indicated. B) Protein concentration, C) vesicle concentration, D) vesicle-to-protein ratio, and E) vesicles per cell
of HS-5 exosomes isolated by UC at 100 000 × g. The x-axis specifies the cell seeding density per 150 cm2 and exosome (Exo) production time. Data
represent mean + SD, n = 3. F) Western blot analysis of exosomes purified by UC at 100 000 × g without microfiltration, for the contaminant calregulin.
Exosomes were produced for G) 24 h or H) 48 h and isolated by UC at 100 000 × g. NTA size distributions (left panels) and TEM images (right panels)
of exosomes. Green arrows mark the main vesicle population with their modal diameter and red arrows indicate vesicle aggregates. I) Cryo-TEM image
of exosomes produced over 48 h and isolated by UC at 100 000 × g. Experiments (panel (F); left panels in (G) and (H)) were performed in triplicates
and a representative image/plot is shown.
enrichment of this EV subtype by both isolation approaches. 2.2. Preparation Conditions Affect the Exosome Characteristics
In contrast to UF-SEC, UC was able to purify enzymati-
cally active HS-5 exosomes in a more reproducible fashion, Selected production and purification parameters were screened
and was hence retained as purification strategy in subsequent to maximize the yield of intact vesicles, while maintaining the
experiments. purity of the exosome preparation (Figure 2A).
The effects of the cell seeding density and exosome produc- 2.3. HS-5 Exosome Characteristics Remain Stable upon Storage
tion time on the vesicle yield were first investigated. Importantly,
the HS-5 cell viability was greater than 95% for all tested con- The stability of HS-5 exosomes was monitored after short- and
ditions, hence ruling out the risk of excessive contamination long-term storage at different temperatures. Incubation for 48
with dead cells, debris or apoptotic bodies. Moreover, 6 × 106 h at 37 °C did neither significantly alter the investigated physi-
cells/150 cm2 was the maximal seeding density investigated as cal properties nor the CD73 activity of the exosomes compared
overconfluence was observed at higher plating densities. Vesicle to 24 h and time zero, verifying that exosome properties re-
yields increased as a function of the cell seeding density until it mained consistent during the 48-h vesicle production period (Fig-
plateaued at a seeding density of 4 × 106 cells/150 cm2 , which cor- ure S3A–D, Supporting Information). Upon long-term storage (6
responded to ≈80% cell confluence at the onset of the exosome months) at −20 and −80 °C, no significant changes were observed
production (Figure 2B,C). Purity of the exosomes produced from in the vesicles’ modal diameter, zeta potential or CD73 enzymatic
different cell numbers was similar, as indicated by the virtually activity (Figure S3B–D, Supporting Information), pleading for a
stable vesicle-to-protein ratio (Figure 2D). In contrast, the nor- preserved exosome integrity. Vesicle counts seemed to increase
malized vesicle yield per cell was highest for a seeding density of after 6 months storage at −80 °C, but not −20 °C, for reasons that
4 × 106 cells/150 cm2 (Figure 2E), rendering it the optimal cell remain to be understood (Figure S3A, Supporting Information).
seeding density maintained in following experiments. Extending
the exosome production to 48 h significantly improved the ex- 2.4. HS-5 Exosome Protein Signature Is Similar between
osome yields compared to 24 h (Figure 2B,C,E). Moreover, the Passages #05 and #15
higher vesicle-to-protein ratio obtained after 48 h indicated a re-
duction of co-isolated protein contaminants (Figure 2D). Similar The exosome composition was monitored as a function of cell
trends were recently observed for primary amniotic fluid hMSC passage to identify the window of passages within which con-
exosomes.[33] sistent exosome properties could be expected. Based on the ob-
Upon evaluation of pre-isolation steps for their impact on the served consistent HS-5 growth kinetics, four independent exo-
vesicle purity, microfiltration was demonstrated to be indispens- some batch replicates were produced for passages #05, #10, and
able for warranting high vesicle purities, albeit it may diminish #15 of the HS-5 cell culture (Figure 3A), and were analyzed by
the vesicle yield.[34,35] This was indicated by the appearance of quantitative proteomics. A total of 2286 proteins were identi-
the contamination marker calregulin when differential low-speed fied, from which 1787 were present in exosomes from cells of
centrifugation on its own was used to process the conditioned all passages (Figure S4A, Supporting Information). Generally,
medium (CM, i.e., cell culture supernatant after the exosome pro- all exosome batches had a similar global protein composition,
duction period) (Figure 2F). whereby the intra-passage similarity was higher than the inter-
Excessive centrifugal forces may compromise the exosome passage one, demonstrating robustness of the exosome batches
morphology (i.e., aggregate formation).[23,36] While this was true independently produced from the same HS-5 cell passage
for UC isolation at 150 000 × g, lower centrifugal forces of (Figure 3B). Similarity was verified by two-group differential pro-
100 000 × g resulted in an increased proportion of single non- tein abundance analysis and p-value distributions comparing pas-
agglomerated vesicles with a smaller modal diameter of 104 ± sages #05 vs #10 and #10 vs #15 (Figure 3C; Figure S4B, Sup-
19 nm (Figure 2G; Figure S1C, Supporting Information). Fur- porting Information). Only three proteins (DNM2, ATP11B, and
thermore, longer centrifugation times of up to 4 h have been GPC4) (Table S1, Supporting Information) were differentially
reported to increase the vesicle yield without impairing their abundant to a significant extent between passages #05 vs #10, but
integrity.[23,37] While the vesicle recovery did not increase propor- not between #10 vs #15 (Figure 3C; Figure S4C, Supporting Infor-
tionally upon doubling the processing time from 70 to 130 min mation). No clear differences were found in the levels of the top
per UC run (1.2 ± 0.6 × 1011 vesicles mL−1 or 97 ± 11 µg pro- eight abundant proteins at the different passages (Figure S4D,
tein mL−1 for 130 min per run vs 6.8 ± 1.0 × 1010 vesicles mL−1 Supporting Information). Albeit a significant downward trend
or 87 ± 18 µg protein mL−1 for 70 min per run), the vesicle-to- was observed in the CD73 protein levels with increasing passage
protein ratio was slightly increased (1.2 ± 0.7 × 109 vesicles µg−1 number, it resulted in no significant reduction of the CD73 ac-
for 130 min per run vs 7.9 ± 0.1 × 108 vesicles µg−1 for 70 min per tivity and was hence not considered as relevant (Figure 3D). Fur-
run), indicating less co-isolated contaminating proteins. Since thermore, in line with the immunoblot analysis, quantitative pro-
these trends were not significant, a run time of 70 min was main- teomics confirmed the presence of exosome markers as well as
tained to not overly prolong the purification process. sample purity (Figure 3E).
In summary, the optimal preparation conditions included a Remarkably, a total number of 144 proteins with a documented
cell seeding density of 4 × 106 /150 cm2 , a 48-h vesicle production role in wound healing were identified in the HS-5-derived ex-
period, a microfiltration step, and a centrifugal force of 100 000 osomes (Figure S4E, Supporting Information). These proteins
× g for 70 min per run. The resulting exosomes eventually had could be classified by their function in extracellular matrix orga-
a modal diameter of 89 ± 7 nm and were not aggregated (Fig- nization, proliferation, migration, epithelialization, collagen and
ure 2H; Figure S1D, Supporting Information). Importantly, the glycosaminoglycan biosynthesis, and angiogenesis (Figure S4E
exosome enrichment and purity were verified by immunoblot- and Table S2, Supporting Information). Importantly, protein lev-
ting (Figure S2, Supporting Information) and the exosome func- els were similar intra- and inter-passage (Figure S4F, Supporting
tionality was corroborated by measuring their intrinsic CD73 ac- Information), suggesting that no significant differences in the bi-
tivity (3.9 ± 0.6 U mg−1 ). Intact vesicle structures were confirmed ological activity of the exosomes produced from different HS-5
by cryo-TEM analysis (Figure 2I). cell passages should be expected.
Figure 3. HS-5 exosome protein profiles are consistent between cell passages #05 and #15. A) Four independent exosome batches each were produced
from HS-5 cells in passages #05, #10, and #15 and analyzed by quantitative proteomics. B) Heat map with hierarchical clustering of inter-sample
distances on the matrix of the overall normalized protein abundance. The color key is expressed in arbitrary units and darker blue colors indicate
differences. n = 4 per passage number. C) Volcano plots of the quantitative protein analysis presented in two-group comparisons (cell passage #05 vs
#10 (left panel) and #10 vs #15 (right panel)). The x-axis shows the effect size expressed as log2 fold change, which was calculated as ratio of the
averaged normalized protein intensities of two passages (left: [average #10]/[average #05]; right: [average #15]/[average #10]). The y-axis represents the
significance expressed as negative decimal logarithm of the adjusted p-value (−log10(padj)). The horizontal line indicates a false discovery rate (FDR)
of 0.05, and the vertical lines indicate an absolute fold change of 1.5. n = 4 per passage number. D) Normalized CD73 abundance (left panel) and CD73
activity (right panel) are shown as a function of cell passages. Data represent mean + SD, n = 3–4. Significance was calculated with an ordinary two-way
ANOVA followed by a post-hoc Tukey’s multiple comparison test, *p < 0.05. E) Exosome marker proteins (black labels) as well as contamination marker
proteins (red labels) are shown as averaged normalized protein abundance from passages #05, #10, and #15. Data represent mean + SD, n = 3 (single
data points present the average of the relative abundance per passage number #05, #10, and #15).
Collectively, exosome protein signatures were shown to be con- CD73 in HS-5 exosome-driven activities relevant for cutaneous
stant between passages #05 and #15, indicating that the vesi- wound healing was studied using in vitro scratch and angiogen-
cles’ biological profiles can be considered consistent within this esis assays on relevant cell culture models (i.e., keratinocytes,
passage range. Moreover, several well-known proteins involved fibroblasts, and endothelial cells). Based on the lipidomic char-
in processes required for wound healing, including extracellular acterization of HS-5 exosomes (Figure S5, Supporting Informa-
matrix proteins and their integrin receptors or growth factors and tion), synthetic exosome-like liposomes (SELL) were designed to
their receptors,[38] were detected in HS-5-derived exosomes, ver- assess the activity of the most abundant HS-5 exosomal lipids.
ifying their potential function in wound healing. HS-5-derived exosomes significantly and dose-dependently
promoted the scratch wound closure of immortalized, but non-
tumorigenic human keratinocytes (HaCaT cells[42] ) compared
2.5. HS-5 Exosomes Have an In Vitro Wound Healing and to the buffer control (Figure 4A). Depleting the CD73 activity
Angiogenic Potential, Which Is Partially Driven by CD73 and by either thermal treatment of the HS-5 exosomes or simulta-
Lipids neous stimulation with the CD73 inhibitor adenosine 5′-(𝛼,𝛽-
methylene)diphosphate (APCP) partially reduced the exosomes’
Primary hMSC exosomes have been previously emphasized to wound closure-promoting activity (Figure 4A; Table S3, Sup-
promote wound healing, tracing their activity to their protein porting Information). This confirmed the involvement of pro-
and nucleic acid content in the first place.[10,39–41] Here, the in- teins in this process, and suggested that thermostable compo-
volvement of exosomal lipids and the abundantly present enzyme nents other than CD73 (e.g., lipids, nucleic acids) contributed
Figure 4. Purified HS-5 exosomes promote scratch wound closure and endothelial cell tube formation in vitro. Scratch wound healing assay using
A) human HaCaT keratinocytes with representative bright-field images, B) primary human fibroblasts, C) immortalized mouse keratinocytes, and D)
HUVEC. Cells were treated with HS-5 exosomes (0, 10, 30, and 90 µg protein mL−1 ), CD73-inhibited HS-5 exosomes (90 µg protein mL−1 , 30 × 10−6 m
APCP), heat-inactivated HS-5 exosomes (90 µg protein mL−1 , 2.5 h, 65 °C), SELL (1.3 × 1011 vesicles mL−1 ) or variations thereof (either without SM,
or DOPS or DOPE), PC-Chol liposomes (1.3 × 1011 vesicles mL−1 ), APCP (30 × 10−6 m), or buffer control. Epidermal growth factor (EGF, 10 or 50 ng
mL−1 ) or vascular endothelial growth factor (VEGF, 20 ng mL−1 ) were used as positive controls. The wound closure at 24 h after scratch wounding was
normalized to the medium control. E) Capillary tube formation assay in HUVEC. Representative bright-field images and quantification of the total tube
length and the number of junctions. HUVEC were treated with HS-5 exosomes (90 µg protein mL−1 ), CD73-inhibited HS-5 exosomes (90 µg protein
mL−1 , 30 × 10−6 m APCP), heat-inactivated HS-5 exosomes (90 µg protein mL−1 , 2.5 h, 65 °C), SELL (1.3 × 1011 vesicles mL−1 ) or variations thereof,
PC-Chol liposomes (1.3 × 1011 vesicles mL−1 ), APCP (30 × 10−6 m), or buffer control or the positive control VEGF (20 ng mL−1 ), and analyzed 20 h
after treatment. A–E) Data represent mean + SD, n = 3–5. Significance was calculated with an ordinary two-way ANOVA followed by a post-hoc Tukey’s
multiple comparison test, *p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001.
as well. On the other hand, SELL had no clear effect on the activity. Thermal treatment, which inactivates thermolabile exo-
wound closure rate (Figure 4A). Similarly, liposomes formu- somal proteins including CD73 (Table S3, Supporting Informa-
lated with 1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC) and tion), significantly diminished the exosome effects to levels ap-
cholesterol (PC-Chol), which are commonly used for drug deliv- proaching those of the exosomal lipids (Figure 4E).
ery purposes,[43,44] were inactive. In the presence of the prolif- Taken together, these results suggested that HS-5 exosomes
eration inhibitor mitomycin C, the HS-5 exosomes’ effects were directly promote the scratch wound closure of keratinocytes, fi-
less pronounced (Figure S6, Supporting Information), indicating broblasts, and endothelial cells, at least partially via CD73 activity.
a predominant pro-mitogenic and a less marked pro-migratory Moreover, HS-5 exosomes showed a direct pro-angiogenic activ-
activity under the tested experimental conditions. These findings ity, which was mainly dependent on exosomal lipids.
revealed that CD73 and exosome components other than lipids
expedited the scratch wound closure in the HaCaT cell model.
HS-5 exosomes further promoted scratch wound closure 2.6. HS-5 Exosomes Show a Promising Therapeutic Potential for
of primary human fibroblasts and immortalized mouse ker- Cutaneous Wound Healing in Healthy Mice
atinocytes, while liposomes irrespective of their composition
were inactive (Figure 4B,C). Moreover, CD73 inhibition partially Since HS-5 exosomes showed a positive effect in cell-based
abrogated the exosomes’ effect on the scratch wound closure of wound closure and angiogenesis assays, their efficacy was fur-
immortalized mouse keratinocytes, reaffirming its pivotal role ther investigated in a pilot full-thickness excisional wound heal-
in this process (Figure 4C; Table S3, Supporting Information). ing study in healthy mice. Exosomes were compared to SELL
HS-5 exosomes also promoted scratch wound closure of primary to distinguish between biological activity stemming from either
human umbilical vein endothelial cells (HUVECs), although to lipids or proteins/nucleic acids carried by the exosomes. HS-5 ex-
a lesser extent compared to their effect on fibroblasts and ker- osomes (15 µg protein or 1.5 × 1011 vesicles), SELL (1.5 × 1011
atinocytes (Figure 4D). Intriguingly, SELL, which were inactive vesicles), or vehicle control were injected intradermally at the
on fibroblasts and keratinocytes, had a similar effect on the wound edges immediately post-wounding and then at day 2 and
wound closure rate in HUVEC as HS-5 exosomes, suggesting day 4 (Figure 5A). Wounds were analyzed at day 5 after injury, a
that specific exosome lipids contained in the liposome formula- time point when re-epithelialization and granulation tissue for-
tion were essential contributors to the HS-5 exosome efficacy, at mation are clearly visible.[47,48]
least in the tested concentration. Confirming that mainly lipids Wound healing in healthy mice is highly optimized and there-
promoted the exosomes’ activity, neither thermal treatment nor fore, a significant enhancement of the overall healing process is
CD73 inhibition showed any detectable effects on the exosome rarely observed.[48] Indeed, histomorphometric analysis of hema-
activity on HUVEC (Figure 4D). A closer investigation of the toxylin/eosin (H&E)-stained wound sections at day 5 showed
lipid species contained in the SELL formulation revealed that similar wound closure rates in the different treatment groups
1,2-dioleoyl-sn-glycero-3-phosphoethanolamine (DOPE), but nei- (Figure 5B; Figure S7A, Supporting Information). While the
ther sphingomyelin (SM) nor 1,2-dioleoyl-sn-glycero-3-phospho- wound epithelium length, which reflects the migration distance,
l-serine (DOPS), significantly contributed to the observed effects, even decreased following either exosome or SELL treatment com-
as indicated by the SELL’ loss-of-activity upon depletion of DOPE pared to the control, no clear differences in the thickness of
(Figure 4D). On the other hand, PC-Chol liposomes were inac- the wound epidermis, re-epithelialization and wound contraction
tive. Notably, the modal diameters of SELL and its variations as were observed (Figure 5C–E; Figure S7B, Supporting Informa-
well as PC-Chol liposomes were similar (Table S4, Supporting tion). These observations were confirmed by semi-quantitative
Information), also to HS-5 exosomes, hence ruling out potential wound scoring (Figure S7C,D, Supporting Information). There
effects originating from differences in the size.[45] was a non-significant trend towards a reduced granulation tis-
Angiogenesis plays a pivotal role in wound healing and pri- sue area in either exosome- or SELL-treated wounds (Figure 5F).
mary hMSC-derived exosomes have a reported pro-angiogenic Ki67-staining of proliferating cells showed a mild, but non-
capability.[38,46] Therefore, the activity of the HS-5-derived exo- significant increase in cellular proliferation in the wound epi-
somes was investigated in an in vitro capillary tube formation dermis following HS-5 exosome vs control treatment, while no
assay using HUVEC. Compared to the control, HS-5 exosomes effect was seen after treatment with SELL (Figure 5G). Nei-
significantly increased the total tube length, which reflects the ther exosomes nor SELL clearly enhanced the cell proliferation
tube formation, and the number of junctions, which represents rate in the granulation tissue. Herovici staining, which distin-
the complexity of the tube network (Figure 4E). Consistent with guishes between young and mature collagen fibers, showed an
the findings from the scratch wound healing assay with HUVEC, increase in the ratio of mature to young collagen fibers within the
SELL were almost as efficient as the exosomes in the tested con- granulation tissue in exosome- and SELL-treated wounds (Fig-
centration. All three lipid species (i.e., DOPS, DOPE, and SM) ure 5H; Figure S7E, Supporting Information), suggesting a po-
seemed to be involved in the SELL’ angiogenic activity as seen tential function of exosomes in collagen maturation during the
by a significant reduction of the total tube length and number wound healing process. Most importantly, both HS-5 exosomes
of junctions following either DOPS or DOPE or SM depletion and SELL equally augmented the total area covered by blood ves-
compared to HS-5 exosomes (Figure 4E). Effects seemed to be sels in the granulation tissue (i.e., area that stained positive for
even stronger on the total length of the capillary tube network as the vascular endothelial cell marker Meca32) compared to the ve-
here, a significant difference was also seen compared to SELL. hicle control (Figure 5I). Interestingly, both the area and mean
PC-Chol liposomes were again inactive. Moreover, CD73 inhibi- blood vessel sizes were only significantly increased at the wound
tion resulted in a non-significant reduction of the HS-5 exosome edge, suggesting a potential local action close to the injection
© 2020 The Authors. Published by Wiley-VCH GmbH

www.advancedscience.com
2002596 (8 of 15)
www.advancedsciencenews.com
Adv. Sci. 2020, 7, 2002596

sites (Figure 5I; Figure S7F, Supporting Information). Similar studies,[26,51,52] could present another possible source for EV loss,
trends were observed for vascular maturation (i.e., blood vessels as only EV larger than the pore size can be efficiently separated
surrounded by 𝛼-smooth muscle action (𝛼-SMA)-expressing vas- from smaller contaminating proteins.[51]
cular smooth muscle cells) (Figure 5I; Figure S7F, Supporting During process optimization, it was found that the cultiva-
Information). The area covered by 𝛼-SMA positive cells outside tion conditions and UC centrifugal force were key for a high
of vessels (i.e., myofibroblasts) did not reveal significant differ- yield and morphological integrity of the exosome formulation.
ences in the different treatment groups (Figure S7G, Supporting Furthermore, only differential UC in combination with microfil-
Information). tration could purify the vesicles efficiently from contaminating
proteins. Moreover, the global protein expression profile of HS-5
exosomes was consistent between passages #05 and #15, which
3. Discussion
would also indicate stable vesicle properties during the produc-
Apart from few exceptions,[37,49] studies investigating currently tion. Finally, physical properties and CD73 activity of HS-5 ex-
available exosome isolation methods provide only insufficient osomes remained generally unchanged upon long-term storage
information on the impact of specific production and/or pu- at −20 and −80 °C, demonstrating good shelf life of the isolated
rification parameters on the vesicles’ characteristics. Moreover, vesicles, as was previously proposed.[53]
in most cases the exosome formulation quality assessment Exploiting the optimized and standardized isolation method,
is limited to the minimally required exosome identification the role of selected exosome components involved in HS-5
characteristics (i.e., size, concentration, purity, identity, and exosome-promoted parameters relevant for cutaneous wound
morphology).[23,30,31] In this work, the optimal conditions for the healing was subsequently investigated.
production and purification of bioactive HS-5 exosomes were sys- Quantitative proteomics verified that proteins involved in pro-
tematically investigated, leading to the development of a stan- cesses required for wound healing were abundantly present in
dardized method that reproducibly yielded functional exosomes HS-5-derived exosomes, among which the transmembrane en-
with high efficiency and purity. In addition, a rapid and cell-free zyme CD73 was not reported yet. CD73, whose beneficial action
enzymatic assay was introduced to monitor the exosome func- in graft-versus-host-disease[12] and cartilage repair[13] was previ-
tionality, and indirectly the vesicle integrity. Quantitative pro- ously emphasized, was found to play a fundamental role in the
teomics allowed to determine the cell passage range delivering exosome-mediated acceleration of the scratch wound closure of
exosomes with a consistent global protein composition. keratinocytes. The loss of activity that was observed upon heat-
Good purification efficiencies have been reported for UF and ing the HS-5 exosomes at 65 °C also suggests that thermal treat-
SEC, either individually or combined.[26–28] However, the head- ment at higher temperatures, which was previously shown to
to-head comparative analysis of UC and UF-SEC clearly identi- not substantially affect the morphology or uptake behavior of the
fied UC as the superior exosome purification approach in terms vesicles,[54] should be carefully examined as it may be detrimen-
of yield, purity, and enzymatic functionality. It is well known that tal to the exosome functionality. Notably, the HaCaT cell model
EV can non-specifically bind to and/or block the UF membranes, used in this work is frequently used to study skin biology for it
potentially leading to an increased rejection of subpore-sized con- retains a normal differentiation capacity, though differing in sev-
taminants and EV loss.[6,50] Moreover, the active shear forces may eral aspects from primary human keratinocytes.[42]
provoke the formation of vesicle and/or contaminating protein In endothelial cells, exosomes and SELL showed compara-
clusters and can even damage the vesicles, including their mem- ble effects on both the scratch wound closure and capillary tube
brane enzymes.[6] This is particularly problematic in view of the formation. Sphingolipids and glycerophospholipids, which were
filtration of larger volumes, such as the one used in this work. the main components of the SELL and were as well abundantly
The volume of clarified CM (CCM) was more than twice the one present in the HS-5 exosomes, were previously shown to promote
used by others[26–28] and could therefore explain the overall poor capillary blood vessel network formation and maturation as well
UF-SEC performance. This CCM volume was in fact required as proliferation/migration of human endothelial cells.[55,56] Here,
to produce a sufficient exosome quantity for downstream anal- a leading role of DOPE in promoting the scratch wound closure
yses, and to determine the vesicle recovery after both isolation and, together with DOPS and SM, tube formation activity was
methods based on comparable volumes. The 70 nm pore-sized shown. The fact that thermal inactivation and/or CD73 inhibi-
column used for SEC, a pore size also employed in previous tion not or only minimally affected the HS-5 exosomes’ activity on
Figure 5. HS-5 exosomes in a dosage of 15 µg are partially effective in skin wound healing. A) Full-thickness excisional wounds in female C57BL/6 mice
were treated with HS-5 exosomes (1.5 × 1011 vesicles or 15 µg protein), SELL (1.5 × 1011 vesicles), or vehicle control. Green arrows indicate the day
of intradermal injection, and red arrows indicate wounding and sacrificing of mice. B–F) Histomorphometric analysis of wounds at day 5 after injury.
B) Representative photomicrographs of H&E-stained wound sections (Es = eschar; WE = wound epidermis/epithelium; D = dermis; GT = granulation
tissue), C) average epithelium length and D) thickness, E) percentage of re-epithelialization, and F) GT area. G) Representative wound sections with
Ki-67 staining of proliferating cells and quantification thereof in the WE and GT. H) Representative Herovici-stained wound sections and quantification
of the mature-to-young collagen ratio. White dashed lines outline the wound area. Outliers were identified by the ROUT method (Q = 1%) and removed
(one outlier per treatment and control group). I) Blood vessel formation and maturation in 5-day wounds. Vessel area is expressed as the percentage of
the total area of the WED, WB, and total GT (GT = WED + WB). Representative wound sections co-stained with Meca32 (red) and 𝛼-SMA (green). Mature
vessels, which are surrounded by vascular smooth muscle cells, show co-staining of both markers. Cell nuclei were counterstained with Hoechst 33342
(blue). C–I) Data represent mean + SD, n = 6–11 wounds from 6 mice per treatment group. Significance was calculated with an unpaired, non-parametric
Mann–Whitney U test, *p < 0.05, ** p < 0.01.
endothelial cells compared to keratinocytes indicated the strong wound healing by HS-5-derived exosomes, providing a solid gate-
interplay between the exosomal component activity and the target way for further in-detail investigations of the underlying molecu-
cell type. lar mechanisms these components trigger in target cells/tissue.
In a pilot experiment performed with healthy mice, HS-5 exo- Additionally, the key conditions of the exosome preparation pro-
somes as well as SELL had no clear effects on the wound closure, cess that critically determine the exosome quality were estab-
wound epidermis thickness, re-epithelialization, or wound con- lished, demonstrating the strong interdependence of the prepa-
traction. This was not unexpected, since the wound healing pro- ration conditions and the exosome formulation quality.
cess in healthy mice is highly optimized.[38,48] Surprisingly, how-
ever, both treatments even reduced the wound epidermis length
and granulation tissue area, but for reasons that remain unclear.
In contrast, HS-5 exosomes but not their liposomal counterparts, 4. Experimental Section
slightly increased the cell proliferation rate at the wound site in Materials: The HPV-16 E6/E7 transformed human bone marrow stro-
healthy mice, suggesting that the dosage applied in vivo might mal cell line HS-5 was purchased from ATCC (Manassas, VA, USA). Pri-
need to be increased to produce significant effects. Most impor- mary human dermal fibroblasts and immortalized human keratinocytes
tantly, both HS-5 exosomes and SELL comparably promoted an- (HaCaT cell line) were kindly provided by Dr. Hans-Dietmar Beer (Depart-
giogenesis, in agreement with the in vitro results. Effects were ment of Dermatology, University Hospital Zurich, Zurich, Switzerland) and
stronger close to the injection sites, suggesting either a mostly Dr. P. Boukamp (Leibniz Institute for Environmental Medicine, Düsseldorf,
Germany), respectively. Mouse keratinocytes were isolated as described
local activity of the exosomes and/or liposomes, or an enhance-
previously.[62] Primary HUVEC were obtained from ScienCell Research
ment of the natural healing process by extending the vascular net- Laboratories (Carlsbad, CA, USA). Defined keratinocyte SFM, Dulbecco’s
work from the wound edge towards the wound bed.[57] Further- Modified Eagle’s Medium (DMEM), EGF, VEGF, fetal bovine serum (FBS),
more, exosomes and SELL favored an increased production of GlutaMAX, l-glutamine, penicillin, streptomycin, HEPES 1 m solution, and
mature collagen fibers, which is expected to result in an early re- phosphate buffered saline (PBS) were obtained from Thermo Fisher Sci-
gain of tensile strength.[58] It remains to be determined whether entific (Waltham, MA, USA). Endothelial basal medium (EBM) was bought
from Lonza Group AG (Basel, Switzerland). APCP, Minimum Essential
the exosomes and/or SELL also affect the scarring response and
Medium Eagle (MEM) Spinner modification, cholera toxin, ethanolamine,
can even limit this effect, as was previously suggested for exo- hydrocortisone, insulin, o-phosphorylethanolamine, and transferrin were
somes produced from adipose-derived hMSC.[59] purchased from Sigma-Aldrich (St. Louis, MO, USA). Sodium chloride
Taken together, these encouraging results suggest that lipids (NaCl), sodium fluoride (NaF), sodium dodecyl sulfate (SDS), sodium
may play an important part in the wound healing activity of ex- orthovanadate (Na3 VO4 ), Tris, sodium citrate, skim milk powder, bovine
osomes, warranting further investigations to more clearly define serum albumin (BSA), 𝛽-mercaptoethanol, Triton X-100 and polysorbate
the dosage at which their effect predominates. The in vivo effects 20, chloroform ReagentPlus, glutaraldehyde, paraformaldehyde, paraffin,
methylcellulose, uranyl acetate, xylene, hematoxylin, eosin, Hoechst 33342
of other exosome components (e.g., CD73 or other proteins with as well as SM (egg, chicken), and cholesterol were also purchased from
a previously reported wound healing promoting activity) may be- Sigma-Aldrich. PureCol bovine collagen solution (3 mg mL−1 ) was ob-
come apparent at higher injected doses, which were not tested tained from Advanced BioMatrix (Carlsbad, CA, USA). Ethylenediaminete-
in this study. Moreover, the dosing regimen could also affect traacetic acid (EDTA), acetone, and acetic acid were obtained from Fluka
the therapeutic outcome. It is believed that exosomes are rapidly Chemie AG (Buchs, Switzerland). cOmplete EDTA-free Protease Inhibitor
cleared from the injection site via the lymphatic system,[60] em- Cocktail was from F. Hoffmann-La Roche Ltd. (Basel, Switzerland). Bro-
mophenol blue was purchased from Alfa Aesar (Haverhill, MA, USA).
phasizing that a more frequent injection or the use of a sus-
Adenosine-5′-monophosphate (AMP), glycerol, and phenylmethylsulfonyl
tainable release formulation and/or topical application of the fluoride (PMSF) were obtained from Acros Organics (Geel, Belgium).
exosomes and subsequent coverage with a dressing could be Malachite green was bought from Bender & Hobein GmbH (Munich,
beneficial.[9,61] It was indeed recently demonstrated that a sus- Germany). Ammonium molybdate tetrahydrate was from abcr GmbH
tained exosome release in the wound bed via a hydrogel formu- (Karlsruhe, Germany). Calcium chloride (CaCl2 ), ethanol, 10 kDa molec-
lation was more effective than a single-dose application, likely ular weight cut-off (MWCO) Amicon Ultra-0.5 mL Centrifugal Filters, and
100 kDa MWCO Centricon Plus-70 Centrifugal Filter Units were obtained
due to a steady-state exosome concentration at the wound site.[9]
from Merck KGaA (Darmstadt, Germany). The lipids DOPC, DOPS, and
Finally, an animal model for impaired wound healing (e.g., di- DOPE (all >99% purity) were obtained from Avanti Polar Lipids (Alabaster,
abetic wounds,[9,48] or use of skin burn injury[10] ) might reveal AL, USA). Primary mouse monoclonal antibodies anti-CD73 (sc-32299),
additional and/or more potent effects of exosomes and SELL, as anti-CD63 (sc-5275), anti-CD9 (sc-13118), anti-TSG101 (sc-7964), anti-
they directly intervened in dysregulated healing processes.[15,46] calregulin (sc-373863), and anti-GAPDH (sc-47724) as well as the West-
Noteworthily, the results in this work have been obtained with ern Blotting Luminol Reagent were purchased from Santa Cruz Biotech-
exosomes derived from the HPV-16 E6/E7 transformed human nology Inc. (Dallas, TX, USA). The secondary HRP-conjugated polyclonal
goat anti-mouse immunoglobulin was obtained from Dako Denmark
bone marrow stromal cell line HS-5, which allowed to obtain A/S (Glostrup, Denmark). Purified rat anti-mouse pan-endothelial cell
large numbers of exosomes with consistent composition and ac- antigen (Meca32) was obtained from BD Biosciences (Franklin Lakes,
tivity. In the future, it will be important to revalidate the results NJ, USA), mouse monoclonal anti-actin 𝛼-smooth muscle-FITC antibody
with primary hMSC as they better mimic a physiological cell from Sigma-Aldrich. Secondary biotinylated anti-rabbit IgG and secondary
behavior. Cy3 AffiniPure donkey anti-mouse IgG antibody were from Jackson Im-
In conclusion, the bioactivity of HS-5 exosomes appeared to munoResearch Europe Ltd. (West Grove, PA, USA). Primary rabbit mon-
oclonal anti-Ki67 (ab15580) antibody was obtained from Abcam (Cam-
be mediated by a broader range of biomacromolecules beyond
bridge, UK). qEVoriginal/70 nm SEC columns were purchased from Izon
yet reported proteins and nucleic acids. In particular, the trans- Science Ltd. (Christchurch, New Zealand). Immun-Blot PVDF membranes
membrane enzyme CD73 and exosomal lipids were shown to with a pore size of 0.2 µm were from Bio-Rad Laboratories Inc. (Hercules,
play an important role in the regulation of processes relevant for CA, USA). Fuji medical X-ray films were purchased from FUJIFILM Europe
GmbH (Düsseldorf, Germany). The 100 nm pore-sized polycarbonate fil- centrifugation at 4 °C and 14 000 × g in 10 kDa MWCO Amicon Ultra-
ter membranes were bought from Sterlitech Corporation (Kent, WA, USA). 0.5 mL Centrifugal Filters to a final volume of ≈120 µL.
Corning 96-well Clear Polystyrene Microplates were purchased from Corn- Protein Quantification: The total protein amount of the exosomes was
ing Inc. (Corning, NY, USA). Tissue culture test plates with 48 wells and determined by the Micro BCA assay according to the manufacturer’s in-
sterile 0.22 µm pore-sized filters were obtained from TPP Techno Plastic structions (Thermo Fisher Scientific). Briefly, 2 µL exosomes were diluted
Products AG (Trasadingen, Switzerland). Carbon-coated grids and 300- with nanopure water to a volume of 150 µL and added per well of a Corn-
mesh lacey carbon-coated copper grids were from Quantifoil Micro Tools ing 96-well Clear Polystyrene Microplate. After addition of 150 µL Micro
GmbH (Grosslöbichau, Germany). Reversed phase nanoEase M/Z Sym- BCA reagent per well, the solutions were incubated at 37 °C for 2 h and the
metry C18 Trap Column (100 Å, 5 µm, 180 µm × 20 mm) and nanoEase absorbance at 562 nm was measured using a Tecan Infinite M200 (Tecan
M/Z C18 HSS T3 Column (100 Å, 1.8 µm, 75 µm × 250 mm) as well as Ac- Group Ltd., Männedorf, Switzerland).
quity UPLC HSS T3 Column (1.8 µm, 150 µm × 50 mm) were obtained CD73 Enzymatic Activity Assay: To determine the CD73 activity of
from Waters Corporation (Milford, MA, USA). Tissue-freezing medium the HS-5 exosomes, a previously published malachite green assay was
was obtained from Leica Biosystems (Wetzlar, Germany). adapted.[64] Briefly, 60 µL containing 0.1 µg exosome protein and 24 µmol
Cell Culture: HS-5 cells, HaCaT, and human fibroblasts were cul- AMP were incubated in a Corning 96-well Clear Polystyrene Microplate
tivated in DMEM supplemented with 100 U mL−1 penicillin, 100 µg for 10 min at RT. To stop the enzymatic reaction, 40 µL color reagent
mL−1 streptomycin, and 10% v/v FBS. Mouse keratinocytes were main- (0.034% w/v malachite green, 1.55% w/v ammonium molybdate tetrahy-
tained in Defined Keratinocyte SFM and MEM in a ratio 2:1 v/v supple- drate, 0.0625% v/v polysorbate 20) was added per well and incubated for
mented with 1.7 µg mL−1 insulin, 3.3 µg mL−1 transferrin, 0.5 µg mL−1 1 h at RT. The absorbance was measured at 620 nm using a Tecan Infinite
o-phosphorylethanolamine, 0.2 µg mL−1 ethanolamine, 0.1 µg mL−1 hy- M200.
drocortisone, 15 × 10−6 m CaCl2 , 0.7 × 10−3 m GlutaMAX, 100 U mL−1 Liposome Preparation: Liposomes were prepared by the thin-film hy-
penicillin, 100 µg mL−1 streptomycin, 67 × 10−12 m cholera toxin, 2.7% dration method.[65] For PC-Chol liposomes, DOPC and cholesterol were
v/v chelated FBS, and 10 ng mL−1 EGF. HUVEC were cultured in EBM dissolved in chloroform ReagentPlus at a molar ratio of 70:30. For
containing 100 U mL−1 penicillin, 100 µg mL−1 streptomycin, 20% v/v SELL, a mixture composed of DOPC, DOPS, DOPE, SM, and choles-
FBS, 2 × 10−3 m l-glutamine, and 10 µg mL−1 hydrocortisone. All cells terol was dissolved in chloroform ReagentPlus at a molar ratio of
were cultured under humidified conditions at 37 °C and 5% CO2 in a Her- 21:14:17.5:17.5:30.[43,66] As indicated in the results section, variations of
acell 240i CO2 Incubator (Thermo Fisher Scientific) and routinely tested to SELL were created by replacing either DOPS, DOPE, or SM by DOPC. Thin
be negative for mycoplasma using the MycoAlert Mycoplasma Detection lipid films were generated by initial solvent removal at 2000 Pa with a rotary
kit according to the manufacturer’s instructions (Lonza Group AG). Cell evaporator (Büchi Labortechnik AG, Flawil, Switzerland) and then dried in
viability was determined by the trypan blue exclusion assay as described vacuo for over 12 h at RT. The thin lipid film was then rehydrated in PBS at
previously.[63] 55 °C to a final total lipid concentration of 10 × 10−3 m. Liposomes were
Production of HS-5-Derived Exosomes: Exosomes were produced ac- immediately subjected to ten freeze–thawing cycles and extruded through
cording to a previously published protocol,[23] which was customized. In two-stacked 100 nm pore-sized polycarbonate filter membranes at 55 °C.
brief, 4 × 106 HS-5 cells were seeded in a volume of 15 mL complete growth Liposome suspensions were stored at 4 °C.
medium per T150 cell culture flask and cultured for 40 h. For exosome NTA: The size profile and concentration of exosomes isolated from
production, cells were washed twice with 6 mL PBS and further cultured similar CCM volumes (see sections “UC” and “UF-SEC”) were analyzed
in 18 mL fresh, serum-free medium for 24 and 48 h, respectively. The CM on a NanoSight NS300 equipped with a CMOS camera and a 488 nm laser
was harvested and centrifuged at 4 °C and 2000 × g for 5 min, immediately source (Malvern Panalytical GmbH, Kassel, Germany) as well as on a Ze-
followed by 10 000 × g for 15 min using a Sorvall RC 6 PLUS centrifuge taView equipped with a CMOS camera and a 405 nm laser source (Parti-
equipped with a SA-600 Fixed Angle Rotor (Thermo Fisher Scientific). The cle Metrix GmbH, Meerbusch, Germany). The ZetaView was additionally
supernatant was filtered through a 0.22 µm pore-sized filter to obtain the used to determine the zeta potential of the vesicles as well as the con-
CCM. The CCM was stored at −20 °C. centration and diameter of the liposomes. Samples were diluted in PBS
UC: UC was performed as described previously.[23] A volume of (1:100–1:1000 for exosomes produced over 24 h and 1:1 000 000 for exo-
396 mL CCM was centrifuged at 4 °C and either 100 000 × g or 150 000 somes produced over 48 h and for liposomes) to a concentration of 107 –
× g for 70 min using an Optima XE-90 ultracentrifuge equipped with a 108 vesicles mL−1 , giving 50–200 vesicles per frame. For the NanoSight,
Type 45 Ti Fixed-Angle Titanium Rotor (Beckman Coulter Life Sciences, a video of 60 s was recorded, with the camera level set to 14, the slider
Indianapolis, IN, USA). Exosome-containing pellets were resuspended in shutter to 1259, and the slider gain to 366. Data was analyzed with the
10 mL PBS, pooled, and subjected to a second ultracentrifugation round NanoSight NTA 3.1 Build 3.1.54 software setting a detection threshold of
under equal conditions. The purified exosome pellet was resuspended in 4. For the ZetaView, the sensitivity was set to 85, the shutter to 150, and
200 µL PBS and stored at −20 °C. In case of CD73 activity determination, the frame rate to 30. Data was analyzed with the ZetaView software ver-
isolated exosomes were resuspended in HEPES buffered saline (HEBS, sion 8.04.04 SP2 applying a bin class width of 5 nm, minimum brightness
20 × 10−3 m HEPES, 150 × 10−3 m NaCl, pH 7.4). of 25, minimum area of 5, maximum area of 1000, and trace length of 15.
UF-SEC: UF-SEC was carried out as described before.[26] A volume Western Blot: Western blot analysis was performed as described
of 360 mL CCM was concentrated to ≈400 µL by centrifugation at 4 °C previously.[67,68] Briefly, 106 HS-5 cells were harvested, centrifuged at 300
and 3500 × g using 100 kDa MWCO Centricon Plus-70 Centrifugal Fil- × g for 5 min, washed with PBS and resuspended in 100 µL 1× lysis buffer
ter Units. The concentrated CCM was applied on qEVoriginal/70 nm SEC (20 × 10−3 m Tris, 150 × 10−3 m NaCl, 5 × 10−3 m EDTA, 1% v/v Triton
columns and 500 µL fractions were collected. Due to the low exosome X-100, 25 × 10−3 m NaF, 1 × 10−3 m PMSF, 1 × 10−3 m Na3 VO4 , cOmplete
amount present per fraction, exosome identification guidelines accord- EDTA-free Protease Inhibitor Cocktail). Cells were lysed on ice for 30 min
ing the International Society for Extracellular Vesicles[30,31] could not be and subsequently centrifuged at 4 °C and 10 000 × g for 10 min to collect
followed. Alternatively, exosome-containing fractions were identified by a the supernatant. Isolated exosomes were diluted 5:1 v/v in 5× lysis buffer
protein/CD73 activity cut-off requiring only very low amounts of exosomal and incubated on ice for 30 min. Then, samples (5–10 µg protein) were
protein. Their protein concentration was determined by the Micro bicin- diluted 5:1 v/v in reducing sample buffer (0.25 m Tris, 10% w/v SDS, 30%
choninic acid (BCA) assay (see section “Protein Quantification”). To de- v/v glycerol, 0.02% w/v bromophenol blue, 5% v/v 𝛽-mercaptoethanol)
termine the CD73 enzymatic activity of the fractions, 3 µL per fraction was and heated at 95 °C for 10 min. Proteins were resolved on 12% SDS poly-
incubated with 40 × 10−6 m AMP in a reaction volume of 60 µL for 20 h at acrylamide gels and transferred on 0.2 µm pore-sized Immun-Blot PVDF
room temperature (RT). AMP hydrolysis was then determined by a mala- membranes. The membrane was blocked with 5% w/v skim milk powder
chite green assay (see section “CD73 Enzymatic Activity Assay”). Fractions dissolved in Tris buffered saline (TBS-T, 20 × 10−3 m Tris, 150 × 10−3 m
with less than 50 µg mL−1 protein and more than 50% AMP hydrolysis NaCl, 1% v/v polysorbate 20) for 2 h at RT. The membrane was probed with
(qEV1/qEV2: fractions 8–13; qEV3: fractions 7–14) were concentrated by primary mouse monoclonal anti-CD73, anti-CD63, anti-CD9, anti-TSG101,
anticalregulin, and anti-GAPDH antibody for overnight at 4 °C. Then, the lyzed with FIJI software applying the Skeletonize3D and AnalyzeSkeleton
membrane was washed four times for 5 min each with TBS-T and incu- plugins.[69,72,73]
bated with the secondary HRP-conjugated polyclonal goat antimouse im- Sample Preparation for Quantitative Proteomics Analysis: For quanti-
munoglobulin for 2 h at RT. The membrane was again washed four times tative proteomics analysis, exosome samples were processed using the
for 5 min each with TBS-T and incubated with Western Blotting Luminol iST Kit according to the manufacturer’s instructions (PreOmics GmbH,
Reagent for 2 min at RT. Protein bands were developed on Fuji medical Planegg/Martinsried, Germany). In brief, exosome samples (20 µg) were
X-ray films in an Agfa Curix 60 (Agfa Corporate, Mortsel, Belgium). diluted in “Lyse” buffer, boiled at 95 °C for 10 min and subsequently pro-
TEM: The exosome morphology was visualized by TEM as described cessed with high intensity focused ultrasound (30 s, 85% amplitude). Sam-
elsewhere.[23] Exosomes (2 µL) were blotted on glow-discharged (30 s ples were loaded in the supplied cartridge and on-filter digested with 50 µL
in an Emitech K100X glow discharge system, Quorum Technologies Ltd., “Digest” solution for 60 min at 37 °C. Then, 100 µL “Stop” solution was
Lewes, UK) carbon-coated grids. Grids were washed with PBS for 2 min, added and the cartridge was centrifuged at 3800 × g. Digested peptides
fixed with 1% v/v glutaraldehyde for 5 min, and rinsed eight times with were washed with 100 µL of each “Wash 1” and “Wash 2” solution, eluted
double-distilled water for 2 min each. Excess liquid was drained off with in 100 µL “Elute” solution, dried, and resuspended in 20 µL “LC-Load”
filter paper. The sample was stained with uranyl oxalate at pH 7 for 5 min buffer for mass spectrometric (MS) analysis.
and then incubated in a mixture of 2% v/v methylcellulose and 4% v/v Liquid Chromatography-Mass Spectrometry (LC-MS) Analysis for Quanti-
uranyl acetate for 10 min on ice. Excess liquid was drained off and the grids tative Proteomics: In a randomized order, 3 µL per peptide sample was
were air-dried. The fixed sample was analyzed in a FEI Morgagni 268 mi- separated on a nanoEase M/Z Symmetry C18 Trap Column coupled to a
croscope (Field Electron and Ion Company, Hillsboro, OR, USA) operated nanoEase M/Z C18 HSS T3 Column by a 0.1% formic acid in water (sol-
at a 100 kV acceleration voltage in the bright field mode. Size distributions vent A) and 0.1% formic acid in acetonitrile (solvent B) gradient (from
from TEM images were obtained with FIJI software.[69] 8% to 22% B in 80 min, 32% B in 10 min, and 95% B in 1 min) at a
Cryo-TEM: Vitrified exosomes were analyzed by cryo-TEM. The ex- flow rate of 0.3 mL min−1 using an M-Class Ultra Performance LC (Waters
osome suspension (3 µL) was added on glow-discharged (30 s in an Corporation). Separated peptides were then analyzed on-line on an Orbi-
Emitech K100X glow discharge system, Quorum Technologies Ltd.) 300- trap Fusion Lumos Tribrid mass spectrometer (Thermo Fisher Scientific)
mesh lacey carbon-coated copper grids. The sample was vitreously frozen equipped with a Digital PicoView nanospray source (New Objective Inc.,
in a mixture of liquid ethane and propane in a Vitrobot Mark II (Field Elec- Woburn, MA, USA). Full-scan MS spectra (300–1500 m/z) at a resolution
tron and Ion Company). Excess sample was removed by controlled blot- of 120 000 at 200 m/z, and a target value of 500 000 charges per acquisition
ting. Grids were mounted in a Gatan cryo-holder and transferred into a Tec- were recorded in the data-dependent acquisition mode. Data-dependent
nai F20 Cryo (Field Electron and Ion Company). The microscope was op- tandem MS spectra were recorded in the linear quadrupole ion trap oper-
erated at a 200 kV acceleration voltage in the bright field mode and main- ated in the rapid scan mode with a target value of 10 000 charges per ac-
tained at −180 °C during sample analysis. Micrographs were recorded un- quisition, a 0.8 Da isolation width, a 50 ms maximum injection time, and
der low-dose conditions (<500 electrons nm−2 ) using a Falcon II 4K Direct a 35% higher energy collision induced dissociation fragmentation energy.
Electron Detector (Field Electron and Ion Company). Only precursor ions with an intensity above 5000 were selected for tandem
In Vitro Scratch Wound Healing Assay: The scratch wound healing as- MS, with the maximum cycle time set to 3 s and an enabled charge state
say was performed as previously reported.[70] Immortalized human ker- screening. Singly, unassigned and charge states greater than seven were
atinocytes (HaCaT cells[42] ) and spontaneously immortalized mouse ker- rejected. Precursor masses previously selected for tandem MS measure-
atinocytes were seeded at a density of 105 000 cells per well and primary ments were excluded from further selection for 30 s, and the exclusion win-
human dermal fibroblasts at 30 000 cells per well in a 48-well plate. Primary dow was set at 10 ppm. Real-time calibration on an internal lock mass of
HUVEC were seeded at 50 000 cells per well in a 48-well plate precoated 371.1012 and 445.1200 m/z was performed. The MS proteomics data was
with PureCol bovine collagen type I (50 µg mL−1 ). Cells were grown to a handled using the local laboratory information management system.[74]
confluent monolayer and a cross-shaped wound was scraped into it using MS Proteomics Data Analysis: The MS raw data was searched against
a sterile 200 µL pipette tip. In some migration studies with HaCaT, cellular the Swissprot Homo sapiens reference proteome (taxonomy 9606, ver-
proliferation was inhibited by incubation with mitomycin C (2 µg mL−1 ) for sion from 2019-07-09) and concatenated to its reversed decoyed fasta
2 h prior to scratching. The cells were washed with PBS and treated with database and common protein contaminants using the integrated An-
HS-5 exosomes (10, 30, or 90 µg protein mL−1 ), heat-inactivated HS-5 exo- dromeda search engine in MaxQuant (version 1.6.2.3).[75] Cysteine car-
somes (90 µg protein mL−1 , 65 °C, 2.5 h), CD73-inhibited HS-5 exosomes bamidomethylation and N-terminal protein acetylation was set as either
(90 µg protein mL−1 , 30 × 10−6 m APCP), SELL (1.3 × 1011 vesicles mL−1 ) fixed or variable modification. Enzyme specificity was set to trypsin/P al-
or variations thereof, PC-Chol liposomes (1.3 × 1011 vesicles mL−1 ), EGF lowing a minimal peptide length of seven amino acids and a maximum
(10 or 50 ng mL−1 ) or VEGF (20 ng mL−1 ), buffer control or APCP con- of two missed-cleavages. A false discovery rate (FDR) of 0.01 for peptides
trol. The concentrations refer to a volume of 150 µL per well. The wound and 0.05 for proteins was used. Label free quantification intensities were
was imaged with a Leica DMI6000 B epifluorescence microscope (Leica log2 transformed and normalized. Statistical analysis was performed us-
Microsystems, Wetzlar, Germany) at a 2.5× magnification after 0 and 24 ing the R environment for statistical computing with Bioconductor (ver-
h. The wound area was analyzed with FIJI software.[69] The percentage of sion 3.6.2).[76] Employing the R package limma (version 3.42.0), differ-
the wound closure was calculated and normalized to the medium control. entially abundant protein analysis was modeled using mixed effect lin-
In Vitro Tube Formation Assay: The capillary tube formation assay was ear regression with empirical Bayes variance estimation.[77] Proteins in-
performed according to a prior published protocol,[71] with minor modifi- volved in wound healing were searched and classified against the AmiGO
cations. Primary HUVEC were seeded at a density of 50 000 cells per well 2 database (version 2.5.13).[78] The protein accession code was based on
in a 48-well plate precoated with PureCol bovine collagen type I (50 µg UniProt.[79]
mL−1 ) and grown to confluency. Cells were washed with PBS and starved Lipidomics Analysis: For lipidomics analysis, exosomes were resus-
for 4 h in EBM-2 (EBM supplemented with 2% v/v FBS, 100 U mL−1 pended in methanol/water 4:1 v/v and centrifuged at 4 °C and 16 000 ×
penicillin, and 100 µg mL−1 streptomycin). HS-5 exosomes (90 µg pro- g for 15 min. Exosomal lipids (1 µL of the supernatant) were separated
tein mL−1 ), heat-inactivated HS-5 exosomes (90 µg protein mL−1 , 65 °C, on a Acquity UPLC HSS T3 Column using a 5 × 10−3 m ammonium ac-
2.5 h), CD73-inhibited HS-5 exosomes (90 µg protein mL−1 , 30 × 10−6 m etate in water/acetonitrile 95:5 (A) and 5 × 10−3 m ammonium acetate in
APCP), SELL (1.3 × 1011 vesicles mL−1 ) or variations thereof, PC-Chol li- isopropanol/acetonitrile 90:10 (B) gradient (from 5% B to 100% B over
posomes (1.3 × 1011 vesicles mL−1 ), VEGF (20 ng mL−1 ), buffer or APCP 12 min) at an adjusted flow rate of 3–4 µL min−1 over 12 min using a
control were suspended in EBM-2 containing 1 mg mL−1 PureCol bovine nanoAcquity UPLC (Waters Corporation). Separated lipids were then an-
collagen type I pH 7.4 and 400 µL were added per well. The tube forma- alyzed on-line on a Q Exactive Mass Spectrometer (Thermo Fisher Sci-
tion was imaged with a Leica DMI6000 B epifluorescence microscope (Le- entific) equipped with a nanoelectrospray ionization source. MS spectra
ica Microsystems) at a 2.5× magnification after 20 h. Images were ana- (50–1200 m/z) were recorded at a MS resolution of 70 000 and MS/MS
resolution of 17 500 using negative and positive polarization and all ion Acknowledgements
fragmentation. Lipid data sets were analyzed with Progenesis QI Software
(Nonlinear Dynamics, Milford, MA, USA). Detected ions were identified The authors gratefully acknowledge the generous funding from the ETH
based on accurate mass, adduct patterns, as well as isotope patterns by Research Grant (ETH-10 16-1). Furthermore, the authors thank Stephan
searching against the Lipid Maps Structure Database[80] with a mass accu- Handschin (ScopeM, ETH Zurich, Switzerland) for performing the EM
racy tolerance of 5 MDa, and the relative abundance of each lipid species measurements, and Dr. Paolo Nanni as well as Dr. Witold Wolski (Func-
was normalized. tional Genomics Center Zurich, ETH Zurich, Switzerland) for perform-
In Vivo Wound Healing Studies: Animal care and experimental proto- ing the quantitative proteomics experiments and providing guidance for
cols had been approved by the local veterinary authorities (Kantonales Vet- the data interpretation. Finally, the authors acknowledge Dr. Endre Laczko
erinäramt, Zurich, Switzerland) and were performed according to Swiss (Functional Genomics Center Zurich, ETH Zurich, Switzerland) for per-
law. Four 5 mm full-thickness excisional wounds were punched in the dor- forming the lipidomics experiments as well as Dr. Michael Prummer
sum of anaesthetized female C57BL/6JRj mice at the age of 9 weeks,[47] and Dr. Lars Bosshard (NEXUS Personalized Health Technologies, ETH
with 6 mice per treatment group. Immediately postwounding and then ev- Zurich, Switzerland) for their invaluable help with the proteomics analysis.
ery second day, 50 µL of HS-5 exosomes (15 µg proteins or 1.5 × 1011 This study was part of the SKINTEGRITY collaborative research program
vesicles), SELL (1.5 × 1011 vesicles), or PBS were applied by two intra- of “University Medicine Zurich.”
dermal injections (25 µL each) thereof at the wound edges. Wounds were
allowed to heal without dressing and collected at day 5 post-wounding for
analysis.
Analysis of Skin Wounds by Histomorphometry and Immunofluorescence Conflict of Interest
Staining: Wounds were excised and either fixed with ethanol/acetic acid The authors declare no conflict of interest.
95:1 v/v overnight or 4% v/v paraformaldehyde in PBS followed by paraffin
embedding, or directly frozen in tissue-freezing medium. Sections (7 µm)
from the middle of the wound were then processed for either histological
or immunofluorescence analysis. Author Contributions
For histological analysis, sections were stained with H&E and accord-
ing to the Herovici stain procedure[81] and subsequently imaged with a J.-C.L. and B.F.H. designed the overall project and wrote the manuscript.
Panoramic 250 Slide Scanner (3D Histech, Budapest, Hungary). B.F.H. performed the in vitro experiments and analyzed all data. J.-C.L.,
For immunohistochemistry analysis, paraffin sections were dewaxed, S.W., M.B.-Y.G., and B.F.H. planned the in vivo study. M.B.-Y.G. conducted
rehydrated in a xylene/ethanol gradient, and incubated in citrate buffer pH the in vivo study and analyzed the in vivo data. S.W. and M.B.-Y.G. provided
6.0 for 1 h at 95 °C to retrieve antigens. Wound sections were blocked with feedback on the manuscript.
PBS containing 12% w/v BSA for 1 h at RT and incubated with the primary
rabbit monoclonal anti-Ki67 antibody, followed by incubation with a bi-
otinylated secondary anti-rabbit antibody. For bright-field analysis, bound
antibodies were detected with the Vectastain ABC kit and the diaminobeni- Keywords
dine peroxidase substrate kit according to the manufacturer´s instructions CD73, exosomes, human bone marrow stromal cell line, lipids, wound
(Vector Laboratories). For immunofluorescence analysis, cryo-wound sec- healing
tions were washed twice with PBS containing 0.1% v/v Triton X-100 and
subsequently fixed with cold acetone. Sections were washed and blocked
with PBS containing 12% w/v BSA for 1 h at RT, followed by incubation Received: July 9, 2020
with primary antibodies (anti-Meca32, antiactin 𝛼-SMA) for overnight at Revised: September 11, 2020
4 °C. Sections were then incubated with the secondary Cy3 antimouse Published online: October 27, 2020
IgG antibody for 30 min at RT and counterstained with Hoechst 33342.
Stained sections were imaged with an Axioskop 2 microscope equipped
with a Plan-Neofluar objective (20×/0.5NA) and photographed with an
Axiocam HRc camera (all from Carl Zeiss Microscopy GmbH, Jena, Ger- [1] C. Théry, L. Zitvogel, S. Amigorena, Nat. Rev. Immunol. 2002, 2, 569.
many). The Axiovision 4.6 software (Carl Zeiss Microscopy GmbH) was [2] M. Colombo, G. Raposo, C. Théry, Annu. Rev. Cell Dev. Biol. 2014, 30,
used for data acquisition. Images were then analyzed using ImagePro Plus 255.
software (Media, Cybernetics Inc., Rockville, MD, USA). [3] I. Bjørge, S. Kim, J. Mano, B. Kalionis, W. Chrzanowski, Biomater. Sci.
Statistical Analysis: Statistical analysis was performed using GraphPad 2018, 6, 60.
Prism (version 8.2.0, GraphPad Software Inc., San Diego, CA, USA). Data [4] D. G. Phinney, M. F. Pittenger, Stem Cells 2017, 35, 851.
was presented as mean + SD, unless stated otherwise. For in vitro cell- [5] G. Van Niel, G. d’Angelo, G. Raposo, J. Mol. Cell Biol. 2018, 19, 213.
based assays (n = 3–5), relative abundance of CD73 (n = 4 per passage [6] L. M. Doyle, M. Z. Wang, Cells 2019, 8, 727.
number) and CD73 activity (n = 3–4 per passage number), mean differ-
[7] B. György, M. E. Hung, X. O. Breakefield, J. N. Leonard, Annu. Rev.
ences between groups were calculated with an ordinary two-way ANOVA
Pharmacol. Toxicol. 2015, 55, 439.
followed by a post-hoc Tukey’s multiple comparison test (𝛼 = 0.05, p-value
[8] A. M. Silva, J. H. Teixeira, M. I. Almeida, R. M. Goncalves, M. A. Bar-
calculated), assuming a normal distribution of the data. For the in vivo
studies, outliers were identified with the ROUT method (Q = 1%) and ex- bosa, S. G. Santos, Eur. J. Pharm. Sci. 2017, 98, 86.
cluded from the analysis. Outliers (one value per treatment and control [9] H. Henriques-Antunes, R. M. Cardoso, A. Zonari, J. Correia, E. C. Leal,
group) were only identified and removed in the Herovici analysis. Treat- A. Jimenez-Balsa, M. M. Lino, A. Barradas, I. Kostic, C. Gomes, J. M.
ments were compared to the control (both n = 6–11 wounds analyzed Karp, E. Carvalho, L. Ferreira, ACS Nano 2019, 13, 8694.
from 6 mice per group) by an unpaired, non-parametric Mann–Whitney U [10] B. Zhang, M. Wang, A. Gong, X. Zhang, X. Wu, Y. Zhu, H. Shi, L. Wu,
test (𝛼 = 0.05, p-value calculated). W. Zhu, H. Qian, Stem Cells 2015, 33, 2158.
[11] C. Lo Sicco, D. Reverberi, C. Balbi, V. Ulivi, E. Principi, L. Pascucci,
P. Becherini, M. C. Bosco, L. Varesio, C. Franzin, M. Pozzobon, R.
Supporting Information Cancedda, R. Tasso, Stem Cells Transl. Med. 2017, 6, 1018.
[12] S. Amarnath, J. E. Foley, D. E. Farthing, R. E. Gress, A. Laurence, M.
Supporting Information is available from the Wiley Online Library or from A. Eckhaus, J. Y. Métais, J. J. Rose, F. T. Hakim, T. C. Felizardo, Stem
the author. Cells 2015, 33, 1200.
[13] S. Zhang, S. J. Chuah, R. C. Lai, J. H. P. Hui, S. K. Lim, W. S. Toh, Gazeau, C. Wilhelm, C. Cellier, O. Clément, G. Rahmi, ACS Nano
Biomaterials 2018, 156, 16. 2018, 12, 9800.
[14] S. Beloribi, E. Ristorcelli, G. Breuzard, F. Silvy, J. Bertrand-Michel, E. [41] M. Wang, C. Wang, M. Chen, Y. Xi, W. Cheng, C. Mao, T. Xu, X. Zhang,
Beraud, A. Verine, D. Lombardo, PLoS One 2012, 7, e47480. C. Lin, W. Gao, ACS Nano 2019, 13, 10279.
[15] D. S. Wijesinghe, in Wound Healing, Tissue Repair, and Regeneration [42] P. Boukamp, R. T. Petrussevska, D. Breitkreutz, J. Hornung, A.
in Diabetes, 1st ed. (Eds: D. Bagchi, A. Das, S. Roy), Elsevier Inc., Markham, N. E. Fusenig, J. Cell Biol. 1988, 106, 761.
Amsterdam 2020, Ch. 10. [43] M. Lu, X. Zhao, H. Xing, Z. Xun, S. Zhu, L. Lang, T. Yang, C. Cai, D.
[16] D. English, D. N. Brindley, S. Spiegel, J. G. Garcia, Biochim. Biophys. Wang, P. Ding, Int. J. Pharm. 2018, 550, 100.
Acta, Mol. Cell Biol. Lipids 2002, 1582, 228. [44] M.-L. Briuglia, C. Rotella, A. McFarlane, D. A. Lamprou, Drug Deliv.
[17] A. M. Campos, E. Maciel, A. S. Moreira, B. Sousa, T. Melo, P. Transl. Res. 2015, 5, 231.
Domingues, L. Curado, B. Antunes, M. R. Domingues, F. Santos, J. [45] E. C. Cho, Q. Zhang, Y. Xia, Nat. Nanotechnol. 2011, 6, 385.
Cell. Physiol. 2016, 231, 1024. [46] P. Hu, Q. Yang, Q. Wang, C. Shi, D. Wang, U. Armato, I. D. Prà, A.
[18] S. W. Ferguson, J. Nguyen, J. Control. Release 2016, 228, 179. Chiarini, Int. J. Burns Trauma 2019, 7, s41038.
[19] K. W. Witwer, E. I. Buzás, L. T. Bemis, A. Bora, C. Lässer, J. Lötvall, [47] P. Hiebert, M. S. Wietecha, M. Cangkrama, E. Haertel, E. Mavrogona-
E. N. Nolte-‘t Hoen, M. G. Piper, S. Sivaraman, J. Skog, J. Extracell. tou, M. Stumpe, H. Steenbock, S. Grossi, H.-D. Beer, P. Angel, Dev.
Vesicles 2013, 2, 20360. Cell 2018, 46, 145.
[20] T. Lener, M. Gimona, L. Aigner, V. Börger, E. Buzas, G. Camussi, N. [48] M. B.-Y. Greenwald, C. Tacconi, M. Jukic, N. Joshi, P. Hiebert, J.
Chaput, D. Chatterjee, F. A. Court, H. A. d. Portillo, J. Extracell. Vesicles Brinckmann, H. Tenor, R. Naef, S. Werner, J. Invest. Dermatol. 2020,
2015, 4, 30087. https://doi.org/10.1016/j.jid.2020.05.111.
[21] G. K. Patel, M. A. Khan, H. Zubair, S. K. Srivastava, M. d. Khushman, [49] M. A. Livshits, E. Khomyakova, E. G. Evtushenko, V. N. Lazarev, N.
S. Singh, A. P. Singh, Sci. Rep. 2019, 9, 5335. A. Kulemin, S. E. Semina, E. V. Generozov, V. M. Govorun, Sci. Rep.
[22] J. Van Deun, P. Mestdagh, R. Sormunen, V. Cocquyt, K. Vermaelen, 2015, 5, 17319.
J. Vandesompele, M. Bracke, O. De Wever, A. Hendrix, J. Extracell. [50] Y. S. Polyakov, A. L. Zydney, J. Membr. Sci. 2013, 434, 106.
Vesicles 2014, 3, 24858. [51] A. N. Böing, E. Van Der Pol, A. E. Grootemaat, F. A. Coumans, A.
[23] C. Théry, S. Amigorena, G. Raposo, A. Clayton, Curr. Protoc. Cell Biol. Sturk, R. Nieuwland, J. Extracell. Vesicles 2014, 3, 23430.
2006, 30, 3.22.1. [52] R. Vogel, F. A. Coumans, R. G. Maltesen, A. N. Böing, K. E. Bonning-
[24] F. Momen-Heravi, L. Balaj, S. Alian, P.-Y. Mantel, A. E. Halleck, A. J. ton, M. L. Broekman, M. F. Broom, E. I. Buzás, G. Christiansen, N.
Trachtenberg, C. E. Soria, S. Oquin, C. M. Bonebreak, E. Saracoglu, J. Hajji, J. Extracell. Vesicles 2016, 5, 31242.
Biol. Chem. 2013, 394, 1253. [53] M. Lee, J.-J. Ban, W. Im, M. Kim, Biotechnol. Bioprocess Eng. 2016, 21,
[25] R. M. Johnstone, M. Adam, J. Hammond, L. Orr, C. Turbide, J. Biol. 299.
Chem. 1987, 262, 9412. [54] E. Schulz, A. Karagianni, M. Koch, G. Fuhrmann, Eur. J. Pharm. Bio-
[26] R. J. Lobb, M. Becker, S. W. Wen, C. S. Wong, A. P. Wiegmans, A. pharm. 2020, 146, 55.
Leimgruber, A. Moller, J. Extracell. Vesicles 2015, 4, 27031. [55] O.-H. Lee, Y.-M. Kim, Y. M. Lee, E.-J. Moon, D.-J. Lee, J.-H. Kim, K.-W.
[27] J. Z. Nordin, Y. Lee, P. Vader, I. Mäger, H. J. Johansson, W. Kim, Y.-G. Kwon, Biochem. Biophys. Res. Commun. 1999, 264, 743.
Heusermann, O. P. Wiklander, M. Hällbrink, Y. Seow, J. J. Bultema, [56] J. W. Choi, D. R. Herr, K. Noguchi, Y. C. Yung, C.-W. Lee, T. Mutoh,
Nanomedicine 2015, 11, 879. M.-E. Lin, S. T. Teo, K. E. Park, A. N. Mosley, Annu. Rev. Pharamcol.
[28] B. J. Benedikter, F. G. Bouwman, T. Vajen, A. C. Heinzmann, G. Grauls, Toxicol. 2010, 50, 157.
E. C. Mariman, E. F. Wouters, P. H. Savelkoul, C. Lopez-Iglesias, R. R. [57] M. Schäfer, S. Werner, Nat. Rev. Mol. Cell Biol. 2008, 9, 628.
Koenen, Sci. Rep. 2017, 7, 15297. [58] Q. T. Smith, in Trauma: Clinical and Biological Aspects, 1st ed. (Ed: S.
[29] E. A. Mol, M.-J. Goumans, P. A. Doevendans, J. P. Sluijter, P. Vader, B. Day), Springer US, New York 1975, Ch. 3.
Nanomedicine 2017, 13, 2061. [59] L. Hu, J. Wang, X. Zhou, Z. Xiong, J. Zhao, R. Yu, F. Huang, H. Zhang,
[30] J. Lötvall, A. F. Hill, F. Hochberg, E. I. Buzás, D. Di Vizio, C. Gardiner, L. Chen, Sci. Rep. 2016, 6, 32993.
Y. S. Gho, I. V. Kurochkin, S. Mathivanan, P. Quesenberry, S. Sahoo, [60] S. Srinivasan, F. O. Vannberg, J. B. Dixon, Sci. Rep. 2016, 6, 24436.
H. Tahara, M. H. Wauben, K. W. Witwer, C. Théry, J. Extracell. Vesicles [61] S. C. Tao, S. C. Guo, M. Li, Q. F. Ke, Y. P. Guo, C. Q. Zhang, Stem Cells
2014, 3, 26913. Transl. Med. 2017, 6, 736.
[31] C. Théry, K. W. Witwer, E. Aikawa, M. J. Alcaraz, J. D. Anderson, R. [62] J. Yang, M. Meyer, A.-K. Müller, F. Böhm, R. Grose, T. Dauwalder,
Andriantsitohaina, A. Antoniou, T. Arab, F. Archer, G. K. Atkin-Smith, F. Verrey, M. Kopf, J. Partanen, W. Bloch, J. Cell Biol. 2010, 188,
J. Extracell. Vesicles 2018, 7, 1535750. 935.
[32] J. Webber, A. Clayton, J. Extracell. Vesicles 2013, 2, 19861. [63] A. Doyle, J. B. Griffiths, Cell and Tissue Culture: Laboratory Procedures
[33] K. Takov, Z. He, H. E. Johnston, J. F. Timms, P. V. Guillot, D. M. Yellon, in Biotechnology, John Wiley & Sons Inc., Chichester, UK 1995.
S. M. Davidson, Basic Res. Cardiol. 2020, 115, 26. [64] K. W. Harder, P. Owen, L. Wong, R. Aebersold, I. Clark-Lewis, F. Jirik,
[34] M. Y. Konoshenko, E. A. Lekchnov, A. V. Vlassov, P. P. Laktionov, Biochem. J. 1994, 298, 395.
BioMed Res. Int. 2018, 2018, 8545347. [65] M. Hope, M. Bally, G. Webb, P. Cullis, Biochim. Biophys. Acta
[35] H. G. Lamparski, A. Metha-Damani, J.-Y. Yao, S. Patel, D.-H. Hsu, C. Biomembr. 1985, 812, 55.
Ruegg, J.-B. Le Pecq, J. Immunol. Methods 2002, 270, 211. [66] T. Skotland, K. Sandvig, A. Llorente, Prog. Lipid Res. 2017, 66, 30.
[36] L. Issman, B. Brenner, Y. Talmon, A. Aharon, PLoS One 2013, 8, [67] U. K. Laemmli, Nature 1970, 227, 680.
e83680. [68] H. Towbin, T. Staehelin, J. Gordon, Proc. Natl. Acad. Sci. USA 1979,
[37] A. Cvjetkovic, J. Lötvall, C. Lässer, J. Extracell. Vesicles 2014, 3, 23111. 76, 4350.
[38] S. A. Eming, P. Martin, M. Tomic-Canic, Sci. Transl. Med. 2014, 6, [69] J. Schindelin, I. Arganda-Carreras, E. Frise, V. Kaynig, M. Longair, T.
265sr6. Pietzsch, S. Preibisch, C. Rueden, S. Saalfeld, B. Schmid, Nat. Meth-
[39] L. Wang, L. Hu, X. Zhou, Z. Xiong, C. Zhang, H. M. Shehada, B. Hu, ods 2012, 9, 676.
J. Song, L. Chen, Sci. Rep. 2017, 7, 13321. [70] C.-C. Liang, A. Y. Park, J.-L. Guan, Nat. Protoc. 2007, 2, 329.
[40] A. K. Silva, S. Perretta, G. Perrod, L. Pidial, V. Lindner, F. Carn, S. [71] Y. Kubota, H. K. Kleinman, G. R. Martin, T. J. Lawley, J. Cell Biol. 1988,
Lemieux, D. Alloyeau, I. Boucenna, P. Menasché, B. Dallemagne, F. 107, 1589.
[72] I. Arganda-Carreras, R. Fernández-González, A. Muñoz-Barrutia, C. Obenchain, A. K. Oleś, H. Pagès, A. Reyes, P. Shannon, G. K. Smyth,
Ortiz-De-Solorzano, Microsc. Res. Tech. 2010, 73, 1019. D. Tenenbaum, L. Waldron, M. Morgan, Nat. Methods 2015, 12, 115.
[73] T.-C. Lee, R. L. Kashyap, C.-N. Chu, CVGIP: Graph. Models Image Pro- [77] M. E. Ritchie, B. Phipson, D. Wu, Y. Hu, C. W. Law, W. Shi, G. K. Smyth,
cess. 1994, 56, 462. Nucleic Acids Res. 2015, 43, e47.
[74] C. Türker, F. Akal, D. Joho, C. Panse, S. Barkow-Oesterreicher, [78] S. Carbon, A. Ireland, C. J. Mungall, S. Shu, B. Marshall, S. Lewis, A.
H. Rehrauer, R. Schlapbach, presented at EDBT 2010, Lausanne, Hub, W. P. W. Group, Bioinformatics 2009, 25, 288.
Switzerland, March 2010. [79] U. Consortium, Nucleic Acids Res. 2019, 47, D506.
[75] J. Cox, M. Mann, Nat. Biotechnol. 2008, 26, 1367. [80] M. Sud, E. Fahy, D. Cotter, A. Brown, E. A. Dennis, C. K. Glass, A. H.
[76] W. Huber, V. J. Carey, R. Gentleman, S. Anders, M. Carlson, B. S. Car- Merrill Jr., R. C. Murphy, C. R. Raetz, D. W. Russell, S. Subramaniam,
valho, H. C. Bravo, S. Davis, L. Gatto, T. Girke, R. Gottardo, F. Hahne, Nucleic Acids Res. 2007, 35, D527.
K. D. Hansen, R. A. Irizarry, M. Lawrence, M. I. Love, J. MacDonald, V. [81] C. Herovici, Rev. Fr. Etud. Clin. Biol. 1963, 8, 88.

Advanced Science - 2020 - Hettich - Exosomes For Wound Healing Purification Optimization and Identification of Bioactive

Uploaded by

Copyright:

Available Formats

Advanced Science - 2020 - Hettich - Exosomes For Wound Healing Purification Optimization and Identification of Bioactive

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Advanced Science - 2020 - Hettich - Exosomes For Wound Healing Purification Optimization and Identification of Bioactive

Uploaded by

Copyright:

Available Formats

FULL PAPER

Exosomes for Wound Healing: Puriﬁcation Optimization

Exosomes are a therapeutically relevant

© 2020 The Authors. Published by Wiley-VCH GmbH

Adv. Sci. 2020, 7, 2002596

You might also like