Journal Pre-proof

Scaffolds with Controlled Release of Pro-Mineralization Exosomes to

Promote Craniofacial Bone Healing without Cell Transplantation

W. Benton Swanson , Zhen Zhang , Kemao Xiu , Ting Gong ,

Miranda Eberle , Ziqi Wang , Peter X. Ma

PII: S1742-7061(20)30576-6
DOI: https://doi.org/10.1016/j.actbio.2020.09.052
Reference: ACTBIO 6968

To appear in: Acta Biomaterialia

Received date: 27 April 2020

Revised date: 10 September 2020
Accepted date: 29 September 2020

Please cite this article as: W. Benton Swanson , Zhen Zhang , Kemao Xiu , Ting Gong ,
Miranda Eberle , Ziqi Wang , Peter X. Ma , Scaffolds with Controlled Release of Pro-Mineralization
Exosomes to Promote Craniofacial Bone Healing without Cell Transplantation, Acta Biomaterialia
(2020), doi: https://doi.org/10.1016/j.actbio.2020.09.052

This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition
of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of
record. This version will undergo additional copyediting, typesetting and review before it is published
in its final form, but we are providing this version to give early visibility of the article. Please note that,
during the production process, errors may be discovered which could affect the content, and all legal
disclaimers that apply to the journal pertain.

© 2020 Published by Elsevier Ltd on behalf of Acta Materialia Inc.

Scaffolds with Controlled Release of Pro-Mineralization Exosomes to Promote Craniofacial Bone
Healing without Cell Transplantation

W. Benton Swanson1†, Zhen Zhang1†^, Kemao Xiu1, Ting Gong1#, Miranda Eberle2, Ziqi Wang3, Peter X.
Ma1,4,5,6*

1
Department of Biologic and Materials Science, School of Dentistry, University of Michigan, Ann Arbor,
USA.
2
Department of Chemistry, University of Michigan, Ann Arbor, USA
3
Department of Mechanical Engineering, College of Engineering, University of Michigan, Ann Arbor, USA
4
Macromolecular Science and Engineering Center, College of Engineering, University of Michigan, Ann
Arbor, USA
5
Department of Biomedical Engineering, College of Engineering and Medical School, University of
Michigan, Ann Arbor, USA
6
Department of Materials Science and Engineering, College of Engineering, University of Michigan, Ann
Arbor, USA

*To whom correspondence should be addressed:

Peter X. Ma, Ph.D., Richard H. Kingery Endowed Collegiate Professor,
Department of Biologic and Materials Sciences
1011 North University Ave., Room 2211
The University of Michigan
Ann Arbor, MI 48109-1078
Tel: (734) 764-2209
Fax: (734) 647-2110
mapx@umich.edu.

† These authors contributed equally to the present work.

^ Current address: Department of Orthopedic Surgery, West China Hospital, West China School of Medicine,
Sichuan University, Chengdu, China
# Current Address: Hospital of Stomatology, Zunyi Medical University, Zunyi, 563000, China
Abstract

Biomimetic bone regeneration methods which demonstrate both clinical and manufacturing feasibility, as

alternatives to autogenic or allogenic bone grafting, remain a challenge to the field of tissue engineering. Here,

we report the pro-osteogenic capacity of exosomes derived from human dental pulp stem cells (hDPSCs) to

facilitate bone marrow stromal cell (BMSC) differentiation and mineralization. To support their delivery, we

engineered a biodegradable polymer delivery platform to improve the encapsulation and the controlled release

of exosomes on a tunable time scale from poly(lactic-co-glycolic acid) (PLGA) and poly(ethylene glycol)

(PEG) triblock copolymer microspheres. Our delivery platform integrates within three-dimensional tissue

engineering scaffolds to enable a straightforward surgical insertion into a mouse calvarial defect. We

demonstrate the osteogenic potential of these functional constructs in vitro and in vivo. Controlled release of

osteogenic hDPSC-derived exosomes facilitates osteogenic differentiation of BMSCs, leading to

mineralization to a degree which is comparable to exogenous administration of the same exosomes in human

and mouse BMSCs. By recruiting endogenous cells to the defects and facilitating their differentiation, the

controlled release of osteogenic exosomes from a tissue engineering scaffold demonstrates accelerated bone

healing in vivo at 8 weeks. Exosomes recapitulate the advantageous properties of mesenchymal

stem/progenitor cells, without manufacturing or immunogenic concerns associated with transplantation of

exogenous cells. This biomaterial platform enables exosome-mediated bone regeneration in an efficacious and

clinically relevant way.

Keywords: tissue engineering, exosome, controlled release, drug delivery, extracellular vesicle

2
1. Introduction

Annually, over $21 billion is spent on the treatment of bone fractures in the United States, impacting about
one million patients annually, where over half experience delayed healing or nonunion [1]. Physical
impairments resulting from bone injury can lead to significant morbidity and socio-economic cost. Critical
sized bone defects which do not heal spontaneously require surgical intervention to prevent potentially
debilitating implications. In these critical-sized defects endogenous cells require exogenous guidance to bridge
a defect and replace substantial bone loss, in the forms of biologic cues and an artificial matrix [2]. The goal
of bone tissue engineering is to design a trajectory by which these large defects heal to restore the tissue’s
function [3, 4]. Bone grafting is considered the current clinical standard for treating large defects, with
limitations [5]. Autografts struggle to provide adequate amounts of bone which can be harvested and lead to
donor site discomfort and morbidity. Allografts have pathogenic concerns, reduced structural integrity from
commercial processing, and ethical concerns with cadaver bone use [6]. Significant progress in the field of
bone tissue engineering has been made in recent years to design bone-grafting alternatives. Transplantation of
mesenchymal stem/progenitor cells has been demonstrated to enhance bone formation [7]. However, the rate
of clinical translation of these therapies is low due to manufacturing and regulatory concerns, scalability,
immunogenicity of grafts and constructs, efficacy, safety, and cost [8]. We aim to design a three-dimensional
tissue engineering construct which recapitulates the advantageous properties of transplanted stem/progenitor
cells while circumventing translational hurdles associated with cell-based therapies, through the local,
controlled release of stem/progenitor cell-derived exosomes (EXOs) and extracellular vesicles (EVs) to guide
bone regeneration.

Cell-derived EXO/EVs, as well as synthetic liposomes, have received considerable consideration for
therapeutic applications across a number of diseases and clinical indications. Exosomes are lipid bilayer-bound
vesicles with diameters in the range of 50-150 nm and expressing a set of specific surface markers [9-11],
which play important roles in autocrine and paracrine signaling [12, 13]. Since the goal of this work was to
develop delivery systems for EXO/EVs instead of study specific pure exosome properties, for description
simplicity we use the term “exosomes (EXOs)” loosely to include EXOs and EVs with size not too far apart
from EXOs throughout this manuscript. EXOs carry proteins and RNA, including microRNAs (miRNAs),
which are characteristic of their donor cell identity and culture microenvironment [14, 15]. Similar to stem
cells, EXOs have shown important therapeutic potentials in cardiac disease, neurogenesis, osteochondral defect
and tooth tissue repair [11, 16-18]. Advantageously, EXOs have a better safety profile than stem cells in terms
of tumorigenicity and immunogenicity [17, 19]. Clinically, exosome administration is limited to intravenous

3
infusion or direct injection. Exosomes circulate the body rapidly and accumulate in the liver and spleen [20],
requiring a high dose to reach therapeutic efficacy. Additionally, there is risk of off-target effects and short
duration of any benefits. In order to realize the advantageous properties of stem cell-derived exosomes, a
delivery platform is required for their efficient therapeutic delivery in a clinically and biologically relevant
manner

Drug delivery systems allow for controlled release kinetics, single administration instead of routine
dosing, reduced circulating drug concentration, and improved patient compliance in the clinical setting [21].
Poly(lactic-co-glycolic acid) (PLGA), a synthetic biodegradable copolymer, is widely used in controlled
release systems because of its tunable degradation rate and safe degradation biproducts [22]. The double
emulsion technique is the most well-known method to encapsulate proteins and hydrophilic small molecules in
polymeric microspheres [23]. An important consideration in developing controlled release platforms for
complex biologics is maintaining their integrity and bioactivity throughout encapsulation and release. Recent
progress in the field of drug delivery has led to the development of controlled release systems for small
molecules, proteins, and plasmids [24-26]. To fill space in a large bony defect, a tissue engineering scaffold
serves as an artificial matrix to retain and organize cells. Previously our lab has developed nanofibrous,
microporous tissue engineering scaffolds from poly(L-lactic acid) (PLLA) [27, 28]. Nanofibers which mimic
the native extracellular matrix are fabricated by a thermally induced phase separation process; interconnected
macropores are made using a sugar sphere porogen. The high surface area and porosity of these three-
dimensional constructs are ideal for functionalizing their surface with a delivery modality [29], yielding a
three-dimensional implantable construct.

The ideal exosome-delivery solution which demonstrates a high administration efficacy would allow
for local and controlled release of exosomes, protecting their bioactivity. We hypothesize that a scaffold with
an integrated delivery system for osteogenesis-inducing exosomes is sufficient to attract resident cells and
cause their differentiation towards endogenous bone repair, without the need for transplanted stem cells. In this
study, a biodegradable platform technology for the controlled release of stem cell-derived exosomes is
developed and validated. Stem/progenitor cells from dental and craniofacial tissue are conveniently obtainable
and have been shown to regenerate mineralized dental tissues [30-33]. We hypothesize that exosomes derived
from dental pulp stem cells are also osteogenic and can be utilized to facilitate bone regeneration. Human
dental pulp stem cell-derived exosomes are delivered using triblock copolymer microspheres immobilized on a
nanofibrous PLLA scaffold to a critical size mouse calvarial bone defect in vivo and result in bone
regeneration, without the need for transplantation of exogenous stem cells.

4
2. Materials and Methods

2.1. Materials

Polyethylene glycol (Mw = 1,000 g/mol), methyl-polyethylene glycol (MW = 1,000 g/mol), α-bromoisobutyryl
bromide (BIBB), polyvinyl alcohol (PVA) (Aldrich); L-lactide (Altasorb); glycolide, stannous octanoate
(Sn(Oct)2), 2-hydroxyethyl methacrylate (HEMA), azobisisobutyronitrile (AIBN) triethylamine (TEA), copper
(I) bromide (CuBr), N,N,N′,N′′,N′′-pentamethyldiethylenetriamine (PMDETA), Span80, ethyl acetate, alkaline
phosphatase staining kit (Sigma); nile blue acrylamide (Polysciences); tetrahydrofuran (THF), diethyl ether, 9-
anthracecenylmethyl methacrylate, hexane, methanol, methylene chloride, isopropanol, Dow Sylgard 184
polydimethylsiloxane, TRIzol, micro BCA protein assay kit, Alexa-Fluor 555 phalloidin, ProLong Gold Anti-
Fade, Alizarin Red S staining solution, conjugated secondary antibody (Fisher Scientific); D-Trehalose, D-
fructose (Oakwood Chemical); mineral oil (Alfa Aesar); deuterated NMR solvents (CDCl3, D2O) (Acros
Organics); chloroform (VWR Chemical); ethanol (Decon Labs); dialysis tubing (3,500 MW cutoff) (Spectrum
Labs); AZ1518 photoresist, SU-8 photoresist (MicroChem), 3,3,3-trifluoropropyltrichlorosilane (Sigma
Aldrich); Electron-microscopy grade paraformaldehyde, electron-microscopy grade glutaraldehyde (EMS
Diasum); Methyl cellulose (TCI America); Dulbecco's Modified Eagle Medium, fetal bovine serum (FBS),
exosome-depleted FBS, penicillin, streptomycin, phosphate buffered saline (PBS), β-glycerophosphate,
ascorbic acid, dexamethasone (Gibco); All-In-One RT MasterMix (Applied Biological Materials Inc.); 7500
Real Time PCR System, PowerUp Sybr Green Master Mix (Applied Biosystems); EpiQuik Whole Cell
Extraction Kit (Epigentek); 4-12% Bis-Tris gels (NuPAGE); PVDF membrane (BioRad); BrightStar ECL
detection kit (Alkali Scientific); Formvar-carbon coated electron microscopy grid (Ted Pella); miRNeasy Mini
Kit (Qiagen).

2.2. Poly(ethylene glycol) and poly(L-lactic-co-glycolic acid) Block Copolymer Synthesis

Triblock PLGA-PEG-PLGA and diblock PEG-PLGA were synthesized from HO-PEG-OH and H3C-PEG-OH
initiators, respectively. Stoichiometric amounts of the monomers: L-lactide and glycolide were added to a
round-bottom flask with commercial PEG (Mw=1,000 g/mol), and Sn(Oct)2 (monomer/SnOct2 = 100). The
system was purged with positive nitrogen pressure for 30 minutes, then heated to 120°C, where the reaction
proceeded by stannous-catalyzed ring opening polymerization (ROP) for two hours [34].

Polymers are abbreviated as PLXGYA-PEG-PLXGYA where X and Y represent percentage of lactide

and glycolide, respectively. In the case of PL85G15-PEG-PL85G15A: 1.25 g HO-PEG-OH, 4.25 g L-lactide, 0.75

5
g glycolide and 112 µL Sn(Oct)2 were added to a round bottom flask with magnetic stirring and positive
nitrogen inflow, then heated to 120°C to melt. The reaction proceeded for two hours then was quenched by
opening to air and cooling. The crude product was dissolved in a minimum amount of chloroform, precipitated
into 5-times excess volume of cold methanol, and collected by suction filtration, twice. 1H NMR (400 MHz,
CDCl3): 𝛿 1.5 ppm, s, sp3 CH3; 3.6 ppm, m, sp3 CH2; 4.8 ppm, m, glycolide sp3 CH2; 5.2 ppm, m, lactide sp3
CH.

2.3. Nile Blue-PLLA Synthesis

Acrylic-end functionalized PLLA was synthesized from hydroxyethyl methacrylate (HEMA) initiator (0.8
mmol, 86 µL) and L-lactide (40 mmol, 5.760 g), with Sn(Oct)2 (112 µL) in a ring opening polymerization
reaction (ROP) at 120°C, in an inert nitrogen environment [35]. After two hours, the resulting HEMA-PLLA
polymer was exposed to air, cooled, then dissolved in 20 mL chloroform and precipitated in 100 mL cold
methanol and the product was collected, repeated twice. 1H NMR (400 MHz, CDCl3): 𝛿 1.6 ppm, s, sp3 CH3;
5.1 ppm, q, sp3 CH; 5.8 ppm, dt, sp2 CH; 6.1 ppm, dd, sp2 CH; 6.4 ppm, dd, sp2 CH.

HEMA-PLLA (1.40 g), nile blue acrylamide (0.012 mmol, 0.005 g), and freshly recrystallized AIBN
(0.06 mmol, 0.0098 g, recrystallized from methanol) were dissolved in 10 mL dioxane at 70°C and allowed to
react overnight. Solvent was removed by rotary evaporation, the resulting residue was dissolved in a minimum
amount of chloroform and precipitated into cold methanol, then collected by suction filtration. This
purification step was repeated until the methanol is no longer visibly colored by unreacted dye (approximately
four times). 1H NMR (400 MHz, CDCl3): 𝛿 1.5 ppm, s, sp3 CH3; 5.1 ppm, q, sp3 CH; 7.5-8.5 ppm, m, sp2 CH.

2.4. 9-Anthracenylmethyl methacrylate-PEG-PLLA Synthesis

Br-PEG-OH was synthesized from HO-PEG-OH (10 g Mw = 1,000 g/mol) dissolved with TEA (2.2 mL) in
anhydrous THF (50 mL), purged under nitrogen for 30 minutes to remove residual oxygen. The reaction was
submerged in an ice bath and α-bromoisobutyryl bromide (1.20 mL) was slowly added dropwise with a
syringe. The solution was stirred for 24 hours and warmed to room temperature gradually as the ice melted.
After 24 hours, solvent was removed by rotary evaporation. The resulting residue was dissolved in a minimum
amount of dichloromethane (DCM), precipitated into cold diethyl ether (5x excess volume), and collected by
suction filtration. The resulting Br-PEG-OH was lyophilized to a white powder. 1H NMR (400 MHz, D2O): 𝛿
3.07 ppm, q, -CH2-Br; 3.5-3.6 ppm, m, -CH2CH2O-.

To make Br-PEG-PLLA, Br-PEG-OH (1.24 g) was combined with L-lactide monomer (5 g) and
Sn(Oct)2 (112 µL) in a round bottom flask, purged with nitrogen for 30 minutes, with stirring. After 30

6
minutes, the mixture was heated to 120°C for two hours, to carry out ring opening polymerization of PLLA
from the terminal hydroxyl group of the PEG macroinitiator. The solid residue was dissolved in a minimum
amount of chloroform, precipitated into cold methanol (5x volume excess), and collected by suction filtration
to be dried in a vacuum chamber. 1H NMR (400 MHz, CDCl3): 𝛿 1.56 ppm, sp3 CH3; 3.1 ppm, q, -CH2-Br;
3.6-3.65 ppm, m, -CH2CH2O-; 4.7-4.8 ppm, m, glycolide sp3 CH2; 5.1-5.2 ppm, m, lactide, sp3 CH.

9-anthracecenylmethyl methacrylate was polymerized from the terminal bromide via atom transfer
radical polymerization (ATRP). Br-PEG-PLLA (2.0 g), 9-anthracenylmethyl methacrylate (0.1 g), and CuBr
(0.04 g) was added to a round bottom flak and purged with nitrogen for one hour. Separately N,N,N′,N′′,N′′-
pentamethyldiethylenetriamine (PMDETA) initiator (200 µL) was dissolved in deionized water (40 mL) and
purged with nitrogen to remove residual oxygen. The PMDETA-water solution was added via a syringe,
through a rubber septum, to the round bottom flask containing the polymer and stirred for 24 hours at room
temperature. The resulting solution was dialyzed against deionized water using 3,500 Mw cutoff dialysis
tubing, changing the water daily for three days, then the solution was frozen to -80°C and lyophilized. 1H
NMR (400 MHz, CDCl3): 𝛿 1.56 ppm, sp3 CH3; 3.6-3.65 ppm, m, -CH2CH2O-; 4.7-4.8 ppm, m, glycolide sp3
CH2; 5.1-5.2 ppm, m, lactide, sp3 CH, 5.6 ppm, d, sp3 CH2; 6.1-6.2 ppm, m, sp2 CH; 6.5, d, sp2 CH2; 7.5-7.7
ppm, dq, sp2 CH; 8.2 ppm, d, sp2 CH.

2.5. Nuclear Magnetic Spectroscopy:

1
H spectra were recorded with an Inova 400 (Varian), operating at 400 MHz at room temperature, using
CDCl3 or D2O as solvent.

2.6. Gel Permeation Chromatography: Polymer samples were dissolved in tetrahydrofuran (THF, 5 mg/mL)
and filtered using a 200-um syringe filter. GPC analysis was performed on a Shimadzu GPC system containing
three columns in series with refractive index and diode array detectors. The GPC was calibrated with narrow
polydispersity polystyrene standards and the molecular weights are reported as polystyrene equivalents.

2.7. Polydimethylsiloxane Microfluidic Device Design and Fabrication: The microfluidic device was drawn to
scale using computer aided design software (AutoCAD 2017, Autodesk). These geometries consist of two in-
line flow focusing junctions fed by inlets for one oil and two aqueous phases. Following each junction is a
straight channel to allow for droplet stabilization. Droplets were collected through a single outlet into an
external reservoir.

7
 Designs were converted to single polygons using K-Layout (0.25.3). Masks were fabricated on a soda
lime glass (90 mm thick) with low reflective chrome, using AZ1518 photoresist on a Heidelberg uPG 501
Mask Maker by direct writing.

Master molds were fabricated by traditional photolithography. SU-8 2050 photoresist was spin-coated
onto a 4” silicon wafer on an interlocked spinner (Solitec) for 60 seconds, to a thickness of 26 µm. Following a
soft bake, the photoresist was exposed through the mask described above, using a MJB45 aligner. The SU-8
was baked post-exposure, then developed using SU-8 developer, and finally rinsed with isopropyl alcohol and
dried with N2. The height and profile of the resulting photoresist was validated using a profilometer (Dektak)
and differential interference contrast microscope (Olympus BX51). Finally, the SU-8 was hard-baked for 15
minutes at 150°C to further crosslink the polymer.

The surface of the master mold was coated with a releasing agent (0.5 mL, 3,3,3-
trifluoropropyltrichlorosilane), then held under vacuum. PDMS (Sylgard 184 kit) was prepared at a
crosslinking density of 17.5% (40.0 g elastomer, 7.0 g curing agent) and mixed well. Prior to casting, the
PDMS was degassed in a benchtop desiccator for 15-30 minutes, until bubbles were gone. The PDMS mixture
was poured onto the coated master mold to desired thickness and degassed again after casting. PDMS was
cured overnight on a hot plate at 100°C.

The cured PDMS was carefully removed from the master mold and cut to individual microfluidic
chips using a razor blade. 1/32” inlet and outlet holes were made using a biopsy punch. PDMS devices were
covalently bound to glass microscope slides by plasma bonding. The PDMS surface was cleaned with Scotch
tape and isopropyl alcohol; glass slides were rinsed with isopropyl alcohol and dried with N2. A Glen Plasma
Cleaner was used to activate PDMS and glass surfaces using a prescribed recipe (Glen recipe 20). After the
process was completed, PDMS devices were immediately aligned and pressed into the glass microscope slides.
Completed devices were cured at 50°C on a hotplate overnight before use.

2.8. Computational Modeling of Microfluidic Device

Simulations of fluid flow in our microfluidic devices were carried out with COMSOL Multiphysics (5.3a,
University of Michigan CAEN license) using a Two-Phase Flow Phase Field Model. AutoCAD designs of
microfluidic devices were imported as dxf files. The single flow functioning junction model contains three
inlets for two solvent phases (discreet aqueous and oil) and one outlet; the dual flow focusing model of the
complete device contained five inlets for three solvent inlets (discrete water, oil, continuous aqueous).
Immiscible fluids were injected into the channels through their respective inlets, and their flows at the inlet

8
were assumed as fully developed laminar flow. Solvent parameters were defined based on PubChem physical
characteristics: viscosity, density, and surface tension; liquid-liquid surface tension coefficients were from
experimental evidence in the literature [36]. Droplet formation was a function of the interface surface tension
coefficient and flow rate ratio.

2.9. Primary Cells and Culture Conditions

Primary human dental pulp stem cells (hDPSCs) were isolated from healthy human third molars, extracted
from patients 16-20 years of age at the University of Michigan School of Dentistry following literature and our
own prior publications [30-33, 37, 38]. The protocol for isolating hDPSCs was performed according to a
protocol approved by the Institutional Review Board at the University of Michigan, and protocol reported
previously[30, 37]. After the first passage, cells were cultured in maintenance media (Dulbecco's Modified
Eagle Medium (α-MEM) containing 10% fetal bovine serum (FBS) and antibiotics (100U/mL penicillin G;
0.1mg/mL streptomycin, GIBCO)). Mouse bone marrow stromal cells (mBMSCs) were obtained from wild-
type male C57BL/6 mice, 6-8 weeks old, by aspiration of the femoral and tibia bone marrow using 26-gauge
syringe needle [39]. mBMSCs were cultured with maintenance media (α-MEM containing 20% FBS and
antibiotics (100U/ml penicillin G; 0.1mg/ml streptomycin)). All cells were cultured at 37℃ under 5% CO2 in a
humidified environment. Media was changed every two days and cells were used at passage 2-5. For three-
dimensional cell culture experiments, the poly(L-lactic acid) nanofibrous scaffolds were sterilized and wet by
70% ethanol for 10 minutes under weak vacuum, and subsequently for 30 mins under ambient pressure. Then
the scaffolds were washed with phosphate buffered saline (PBS) three times and twice with maintenance
media.

2.10. Osteoinductive Culture and Exosome Isolation by Ultracentrifugation

Human DPSCs were cultured in osteoinductive media (OM) for 14 days. OM consists of -MEM media with
10% FBS, 100 U/mL penicillin, 100 µg/mL streptomycin, and osteoinductive supplements (100 nM
dexamethasone, 10 mM β-glycerophosphate, and 50 µg/mL ascorbic acid). Their mineralization status was
confirmed by Alizarin Red staining. One million DPSCs were initially seeded in a 100mm Petri dish with 10
ml media and grown to about 3-4 million cells when reaching confluency. The culture media was changed
every other day. More exosomes were produced at the mineralizing stage. From the 15th day, the media was
changed to exosome-free osteogenic media (EXO-free OM, OM with exosome-depleted FBS (Gibco)). The
conditioned media (CM) was collected every two days until the 28th day.

9
 hDPSC-derived osteoinductive exosomes (OS-EXOs) were isolated from CM by differential
centrifugation as reported in the literature [40]. Briefly, the CM was centrifuged at 300 g for 10 minutes, 2,000
g for 10 minutes, then 10,000 g for 30 minutes to remove cells and large debris. Then the CM was filtered
through a 0.22-μm filter (GVS North America) to remove smaller residual debris. The supernatant was
concentrated at 4,000 g in a 15 mL Amicon Ultra-15 Centrifugal Filter Unit with a molecular weight cut off
(MWCO) of 100kD (Millipore Sigma) for 10 minutes. Subsequently the CM was ultra-centrifuged at 100,000
g for 70 minutes. Finally, the pellet was resuspended and centrifuged at 100,000 g for another 70 minutes. The
pellet was resuspended with 200 µL PBS and the concentration of exosomes was determined by microBCA
protein assay kit (Thermo Fisher Scientific) before storage at -80℃ until use. About 300 µg exosomes were
obtained at the end from the initial one million hDPSCs, where the osteoinductive supplements were depleted
during the repeated centrifuging and washing with PBS.

2.11. Exosome Physical Characterization by Nanoparticle Tracking Analysis

Average hydrodynamic diameter and concentration of exosomes in solution was characterized by nanoparticle
tracking analysis (NTA) using a NanoSight NS3000 (Malvern). Samples were agitated briefly in a gentle sonic
bath to break up aggregates, were diluted in particle-free PBS and immediately analyzed. NTA was carried out
using a 488 nm laser in scatter mode, using a syringe pump (Harvard Apparatus) to control fluid flow through
the manufacturer’s flow plate. Five, 60-second videos were recorded at a concentration sufficient to obtain a
minimum of 200 tracks per video and averaged for each sample analyzed. Analysis was carried out using
Malvern NanoSight NTA software v3.2. Four independent samples from each time point were measured.

2.12. Fluorescent Labelling and Visualization of Cellular Uptake of Exosomes

Exosomes were labelled with DiO Green (Thermo Fisher Scientific) using a modified membrane-labelling
method reported in the literature [41]. Briefly, 10 µg/mL OS-EXOs were incubated with DiO reagent for 20
minutes, condensed using Amicon 100kD filter, washed with PBS and ultracentrifuged three times. 50,000
mBMSCs were seeded on a glass cover slide and cultured in a 24-well-plate. An aliquot of 8 µg labeled OS-
EXOs was added to 800 µl culture media and incubated for 30 minutes. After washing with PBS, Alexa-Fluor
555 phalloidin was used to stain the cytoskeleton (F-actin, 1:35; 1% BSA), per the manufacturer’s guidelines.
Mounting media containing DAPI was used to stain nuclei (ProLong Gold Anti-Fade). Constructs were
observed using confocal microscopy (Nikon Eclipse C1).

2.13. Alkaline Phosphatase Assay

10
The alkaline phosphatase assay was conducted with a staining kit (Sigma-Aldrich) according to
manufacturer’s instructions. Briefly, the cells were washed with PBS three times before staining. Cells were
fixed by alkaline phosphatase fixative (66% acetone, 25.6% citrate buffer, 8.4% formaldehyde) for 30 seconds,
and the fixative was decanted. Cells were then washed twice by deionized water and incubated with staining
solution for 15 minutes in the dark. Finally, cells were washed by deionized water twice and imaged using an
Olympus light microscope with digital camera.

2.14. Alizarin Red Assay

The cells were washed by PBS 3 times before staining, fixed with 10% formalin for ten minutes, and washed
by deionized water three times. Cells were stained with Alizarin Red S staining solution (20 g/L, pH 4.2) for
15 minutes. After staining the cells were washed by deionized water three times and imaged using an Olympus
light microscope with digital camera.

2.15. Gene Expression Analysis by Polymerase Chain Reaction

Total RNA was isolated using Trizol Reagent (Thermo Fisher Scientific), following a standard protocol [42].
The RNA was reverse transcribed using 5X All-In-One RT MasterMix (Applied Biological Materials Inc.) and
subsequently amplified using 7500 Real Time PCR System (Applied Biosystems) with PowerUp Sybr Green
Master Mix (Applied Biosystems), and primers as shown in Table 1. Relative mRNA expression levels were
normalized with respect to glyceraldehyde-3-phosphate dehydrogenase (GAPDH) mRNA, and expression is
calculated according to ΔΔCt method [43].

2.16. Western Blot Analysis

Western blot analysis was performed according to standard protocol [44]. Briefly, exosomes and cells were
lysed, and proteins were extracted by EpiQuik Whole Cell Extraction Kit (Epigentek) following
manufacturer’s instruction, and the protein concentration was determined by micro BCA protein assay kit
(Thermo Fisher Scientific). 30 µg total protein was denatured and then separated on 4-12% Bis-Tris gels
(NuPAGE). Then the proteins were transferred from gel to a polyvinylidene difluoride (PVDF) membrane and
blocked by 5% BSA solution. After that membrane was incubated with primary antibody in 5% BSA solution
at 4℃ overnight and washed with Tris-buffered saline with 0.1% Tween 20 (TBST) three times, 5 minutes
each. Then the membrane was incubated with a conjugated secondary antibody in 5% BSA solution at room
temperature for one hour, and washed by TBST for three times, five minutes per wash. The signal
development was done with BrightStar ECL detection kit (Alkali Scientific) and images were acquired after
developing in a darkroom.

11
2.17. EXO-MS Fabrication by a Microfluidic Device

Fluids were injected from reservoirs using three Fluigent Flow EZ pressure-driven flow controllers (2,000
mbar). Reservoirs were connected to the microfluidic chip using 1/32” FEP (fluorinated ethylene-propylene
copolymer) solvent-resistant tubing (Cole Parmer). The PDMS device was mounted on a binocular microscope
(AmScope) to visualize droplet formation. Videos and photos were recorded using a 5-megapixel USB camera
(AmScope).

Exosomes isolated from DPSC conditioned media were isolated as described and concentrated to at
least 2,000 µg/mL in PBS. The amount of OS-EXO solution was calculated assuming 85% encapsulation
efficiency based on experience, and dose to match 1 µg/mL treatment regimen when matched for cell number.
D-Trehalose, a hydrophilic sugar which acts as a stabilizer, was added to make a 2 mM solution and was the
discrete aqueous phase. PLGA-b-PEG-b-PLGA copolymer was dissolved in ethyl acetate (2% wt/v) as the
discrete oil phase. 0.01% (wt/wt) polyvinyl alcohol (PVA) in MilliQ water was the continuous aqueous phase.
Solutions were flowed into the PDMS microfluidic device as described above, with precise control of their
pressures (Fluigent FlowEZ). FEP tubing from the outlet flew into a 25 mL round bottom flask which contains
0.01% (w/w) PVA; resulting droplets were collected and stirred with a magnetic stir bar. After 10 mg particles
were collected (determined by amount of discrete polymer phase consumed), particles were transferred to a
beaker in a fume hood, with stirring, where organic solvent was allowed to slowly evaporate. Particles were
collected by centrifugation (15,000 rpm, 12 minutes) and washed with MilliQ water five times before being
frozen at -80°C and lyophilized.

2.18. Scanning Electron Microscopy Characterization

Particles and scaffolds were fixed to specimen mounts by carbon tape. In preparation for observation by
scanning electron microscopy (SEM), samples were sputter-coated with gold (120s, DeskII, Denton Vacuum)
and observed at 5 kV and a working distance of 10 mm.

2.19. Polymer Sphere Size Characterization

EXO-MS particles fabricated via our microfluidic device were measured by laser diffraction using a particle
analyzer (Malvern Mastersizer 2000 with Hydro 2000S wet dispersion unit). Particles were suspended in
MilliQ water (2-5 mg/mL) and added to the sample reservoir, with stirring and sonication. Standard refractive
indices of 1.590 and 1.330 were used for PLGA (particles) and water, respectively. Samples were measured in
triplicate.

2.20. Fabrication of Nanofibrous Tissue Engineering Scaffold

12
Nanofibrous, microporous three-dimensional tissue engineering scaffolds made of poly(L-lactic acid) (PLLA)
were fabricated as previously described in the literature [27]. Briefly, a sugar sphere template was prepared by
emulsifying molten D-fructose in hot mineral oil, with Span80 surfactant. After emulsifying, the mixture was
quenched and mineral oil was exchanged for hexane, sugar spheres were collected. Spheres were separated by
size by molecular sieves, packed into Teflon molds and heat treated at 37°C for 10 minutes to anneal the
spheres. Hexane was removed under vacuum, and polymer solution (PLLA, 10% wt/v in THF, 60°C) was cast
into the mold. The cast polymer scaffolds were immediately frozen at -80°C for a minimum of 48 hours to
induce phase separation, responsible for nanofiber formation. After 48 hours, THF was exchanged for hexane
at room temperature overnight, then scaffolds were removed from molds and sugar was leached in deionized
water. Scaffolds were cut to size (8 mm diameter x 1.5 mm height for subcutaneous implantation, 2.7 mm
diameter x 1 mm height for calvarial defect), lyophilized, and sterilized with ethylene oxide gas prior to in
vitro and in vivo experiments.

2.21. Solvent-wetting Method of Attaching EXO-MS to Nanofibrous Scaffolds

Exosome-releasing spheres (EXO-MS) were physically attached to the surface of nanofibrous, microporous
tissue engineering scaffolds by a post-seeding method [45]. Scaffolds and particles were fabricated separately
by the protocols described above. 1.2 mg particles were suspended in 500 µL hexane with 0.1% Span80, kept
in suspension on a rocker for 30 minutes, then dried in a vacuum chamber. Particles were resuspended in 150
µL a 97:3 hexane:THF mixture which facilitates physical attachment between the PLGA particles and PLLA
scaffold, both of which are soluble in THF, but not hexane. A low amount of THF was used to prevent loss of
the nanofibrous surface structure of the scaffold or disrupt the particle morphology. 15 µL of suspended
particles were seeded to the scaffold, rocked for 30 minutes at a low speed, at room temperature. This seeding
procedure was repeated until a total of 1.0 mg particles were suspended evenly on both sides of the scaffold.
Scaffolds were kept in a vacuum chamber to dry until a consistent mass was achieved.

2.22. Sterilization of Biomaterial Constructs

Prior to further in vitro or in vivo testing, all biomaterial constructs were sterilized using ethylene oxide
(Andersen).

2.23. Evaluation of Controlled Release of Exosomes

The controlled release of exosomes from biomaterial constructs was evaluated in phosphate buffered saline
(PBS, 0.1 M, pH 7.4). 1.0 mg of EXO-MS, or a scaffold seeded with 1.0 mg of EXO-MS, were incubated in
1.0 mL PBS and shaken at 10 rpm in a 37°C incubator. At designated time points (2 h, 4 h, 8 h, 12 h, 24 h, 48

13
h, 72 h, 96 h, 7 d, 10 d, 14 d, 3 w, 4 w, 5 w, 6 w, 7 w, 8 w, 9 w, 10 w, 11 w, 12 w) the PBS was collected,
frozen at -80°C until subsequent analysis by NTA, and replaced with fresh PBS. EXO release was plotted as
cumulative release as a function of time (n=4 per time point).

2.24. Transmission Electron Microscopy

Whole-mount exosome morphology and membrane integrity was evaluated by transmission electron
microscopy. Exosome samples were prepared according to a published protocol [46, 47]. Briefly, EXO-MS
eluent containing OS-EXOs were mixed with an equal volume of 4% electron microscopy-grade
paraformaldehyde and 8 µL was incubated on a Formvar-carbon coated grid for 20 minutes. Grids were
washed with PBS and transferred to a clean surface, then fixed with 50 µL of 1% EM-grade glutaraldehyde.
Grids were washed eight times in fresh deionized water and stained with 1% uranyl acetate solution (pH 3.5)
for 3 minutes and stained with 2% wt/v methyl cellulose for 5 minutes. Excess solution was removed with a
piece of filter paper. Grids were imaged within 24 hours of staining. Samples were observed on a JEOL JEM
1400 TEM at 80.0 kV and 20,000x magnification.

2.25. RNA Isolation from Released Exosomes

EXO-MS eluent from release profile experiments was concentrated at 12,000 g for 90 minutes and 4°C. RNA
was extracted from released exosomes using a phenol-based method [48]. TRIzol was used to extract total
RNA, in accordance with the manufacturers protocol. RNA concentration was measured by absorbance at 260
nm using a Beckman DU640 spectrophotometer.

2.26. Infrared Spectroscopy of 3D Constructs

Three dimensional nanofibrous PLLA scaffolds were functionalized with EXO-MS, cultured with mBMSCs
for up to three weeks, and fixed with 4% PFA. Fourier-transformed mid-infrared (FTIR) spectra were acquired
using a Perkin-Elmer Spectrum GX, with microscope accessory, using a KBr beam-splitter and liquid
nitrogen-cooled MCT-B detector, which covers the range from 4,000 cm-1 to 700 cm-1. The following
parameters were used for each acquisition--resolution: 4 cm-1, J-step: 4 cm-1, interval: 1.00 cm-1, 50 scans per
site. Spectra were collected from ten representative 625 µm2 areas of the construct surface and averaged for
each sample.

2.27. Colorimetric Mineralization Assay

The colorimetric mineralization assay was conducted by the modified o-cresolphthalein complexone method.
Briefly, the cell/scaffold constructs were washed by PBS three times for 5 minutes each, then homogenized in
250 µL 0.5 N Hydrochloric acid. The calcium extraction was performed by shaking overnight at 4 ℃. The

14
total calcium content was determined by calcium assay kit (Pointe Scientific Inc) following the manufactures’
instructions.

2.28. Energy-dispersive X-ray Spectroscopy (EDX) of 3D Constructs

Three dimensional constructs cultured for up to three weeks were fixed with 4% PFA and sputter-coated with
gold (120s, DeskII, Denton Vacuum). Samples were observed using a JEOL JSM-7800FLV scanning electron
microscope equipped with an Oxford XMaxN 80 mm2 silicon-drift energy-dispersive X-ray spectrometer. Ten
spectra were acquired from representative locations in each sample using Oxford Aztec v3.3 EDS acquisition
and processing software and averaged. The same samples were used for FTIR and EDX analysis.

2.29. Subcutaneous Implantation in Mice

200,000 primary mBMSCs were seeded on nanofibrous scaffolds functionalized with OS-EXO loaded EXO-
MS, blank EXO-MS, or left virgin. Constructs were cultured in maintenance media for 24 hours before
subcutaneous implantation in 8- to 10-week-old C57BL/6 mice (Charles River) following a protocol approved
by the University of Michigan Institutional Animal Care and Use Committee (IACUC). Briefly, the mice were
randomly divided into three groups.

Mice were anaesthetized by inhalation of isoflurane (1-5%). One midsagittal incision was made on the
dorsa of each mouse. Two subcutaneous pockets were made by blunt dissection, one on each side. In total four
cell-scaffold constructs were implanted into one mouse with two cell-scaffold constructs placed cephalad and
caudad in each pocket. The incisions were closed with surgical staples and animals were given carprofen to
manage pain. The mice were euthanized by CO2 asphyxiation followed by bilateral pneumothorax 8-weeks
post operation. The implants were dissected and prepared for analysis. For histology, samples were fixed in
4% PFA prior to paraffin embedding and serial sectioning. For gene expression analysis, samples were
subjected to total RNA extraction in TRIzol.

2.30. Calvarial Defect in Mice

All procedures were approved by University of Michigan IACUC. Briefly, 8- to 10-week-old C57BL/6 mice
(Charles River) were allocated into 3 groups randomly. The mice were anaesthetized by inhalation of
isoflurane (1-5%), and a midsagittal incision was made on the scalp, followed by blunt dissection to expose the
parietal calvarial bone. A 2.7 mm craniotomy was performed using a trephine burr centered on the parietal
calvarial bone and suture. A sterilized three-dimensional construct, without cells, was placed to fill the defect,
and the incision was then closed with 4-0 sutures (Ethilon) and animals were given carprofen to manage pain.

15
The mice were euthanized by CO2 asphyxiation followed by bilateral pneumothorax 8-weeks post operation;
calvaria were harvested for further micro-CT and histological studies.

2.31. Microcomputed Tomography Analysis of Bone Formation

For the analysis of calvarial bone, a fixed global threshold of 180 was applied to the microcomputed
tomography (µCT) scan. 3D reconstruction of the skull and quantification of the newly formed bone was
performed. A 2.7 mm round region centered on the defect was determined and the bone volume (BV) and
bone mineral density (BMD) in the area were measured using Scanco software (Scanco μCT 100).

2.32. Histological Analyses

Samples were fixed in 4% PFA and demineralized (calvaria only). After dehydration step, samples were
embedded in paraffin and serial sections of 5 µm thickness were cut. Sections were prepared according to
standard protocols for: hematoxylin and eosin (H&E), Masson’s trichrome, and von Kossa staining. All
sections were evaluated using a light microscope and camera.

2.33. Statistical Methods

All data are reported as mean ± standard deviation. To determine statistical significance of observed
differences between experimental groups, the Student’s t-test was used. p < 0.05 is used to determine
significance. All analysis was carried out in IBM SPSS v23.0 and GraphPad Prism v8.

3. Results

3.1. Isolation of Exosomes from Human Dental Pulp Stem Cells

Human dental pulp stem cells (hDPSCs) are multipotent and can be differentiated towards adipogenic,
osteogenic, and chondrogenic fates [30, 33, 38, 49]. Primary hDPSCs were cultured under osteogenic
conditions for two weeks, then their osteogenic exosomes (OS-EXOs) were isolated by ultracentrifugation
following literature [11, 40]. The physical and molecular identity of these exosomes was assessed to determine
the purity of the exosome isolate following literature and our previous publication [11, 46, 50]. Based on
nanoparticle tracking analysis (NTA), particles in the exosome fraction had an average hydrodynamic diameter
of about 135 nm (Fig 1A, B) with a few minor shoulders of extracellular vesicles outside the size range of the
exosomes. Western blot analysis demonstrated that the OS-EXO fraction was highly enriched for characteristic
CD9, Alix, and Tsg101 proteins (Fig 1C), which were much less enriched in the total cell lysate. Such EXOs
were also shown to strongly express CD63 and CD81 [11]. β-actin was highly enriched in the cell lysate, but

16
not the exosome lysate (Fig 1C). Freshly isolated OS-EXOs were labelled using a lipophilic dye (DiO) for 30
minutes and shown to be rapidly up taken by recipient primary mouse bone marrow stem cells (mBMSCs)
using confocal microscopy (Fig 1D). OS-EXOs were up taken and localized with the recipient cell cytoplasm.

3.2. Characterization of the Osteogenic Potential of hDPSC-derived OS-EXOs

We hypothesized that exosomes derived from primary hDPSCs under osteogenic, or mineralization
conditions would be a potent tool to modulate the mineralization and osteogenic differentiation of recipient
primary mBMSCs. Primary mBMSCs were treated with 1 µg/mL and 5 µg/mL OS-EXOs, compared to
osteogenic and growth media alone for up to two weeks. As early as three days in culture, treatment with OS-
EXOs increased ALP activity compared to osteogenic and growth media (Fig 2A, 2B); ALP activity was
consistently higher in OS-EXO treated cells during the seven-day culture. At 14 and 21 days in culture,
exogenous administration of OS-EXOs increased mBMSC mineralization; an increased dose of OS-EXOs
resulted in higher mineralization, particularly at 14 days (Fig 2C, 2D). Our data suggests that OS-EXOs may
accelerate in vivo bone healing in a dose-dependent manner. The similar therapeutic benefit of OS-EXOs was
shown in primary human BMSCs in vitro at 14 and 21 days (Sup Fig 1).

Increases in mineralization were accompanied by upregulation in the expression of osteogenic

markers including osteocalcin (OCN), bone sialoprotein (BSP) and RUNX Family Transcription Factor 2
(RUNX2) (Fig 2E), and increased protein secretion of the same gene products (Fig 2F, 2G, 2H).

Treatment with exogenous OS-EXOs increased phosphorylation of Erk in a dose-dependent manner

and upregulated Runx2 after 30 minutes (Fig 3A&B). mBMSCs were treated with U1026, an inhibitor of Erk
phosphorylation to probe the involvement of Erk phosphorylation in OS-EXOs mediated Runx2 upregulation
[51]. In control experiments, without treatment by OS-EXOs, U1026 treatment suppressed p-Erk and Runx2
compared with untreated control. Treatment by 5 µg/mL OS-EXOs upregulated p-Erk and Runx2 protein
levels; when treated with U1026 expression of p-Erk with Runx2 decreased simultaneously (Fig 3A&C). In
contrast, OS-EXO treatment did not alter Smad1 phosphorylation (Sup Fig 2). Based on observed changes in
Runx2 and p-Erk translation upon treatment with OS-EXOs alone or in combination with U1026, our results
implicated that OS-EXOs induced osteogenic protein expression through an Erk-dependent way.

3.3. Droplet-based Microfluidics facilitating EXO-MS Formation by Self-Assembly

Previously our group and others have used double emulsion methods to encapsulate small molecules,
proteins and miRNA in poly(aliphatic ester) including poly(L-lactic-co-glycolic acid) (PLGA) microspheres,
which degrade by hydrolysis [23, 26, 39, 52]. There were significant challenges to encapsulate large biologic

17
species such as exosomes using such platform, where block copolymers were needed [53]. Here, droplet-based
microfluidics was used to further facilitate EXO-MS fabrication as shown in the scheme (Fig 4A) and the
microfluidic devices were fabricated from polydimethylsiloxane (PDMS) using soft-lithography methods. The
microfluidic device was designed with three inlets, for each of the discrete cargo-containing aqueous phase,
discrete polymer-containing organic phase, and continuous aqueous phase (Fig 4B, Sup. Fig. 3). The
immiscible water and organic phases contacted at each of two flow-focusing junctions (Fig 4C).

COMSOL fluid dynamics simulations using a phase field method demonstrated droplet formation at
each junction, controlled by the interfacial surface tension between immiscible aqueous and organic phases at
each of the two flow focusing junctions (Fig 4D). The resulting droplets were the product of a
water/organic/water (w/o/w) emulsion and were discrete compartments where the aqueous phases were
physically separated by an organic, polymer-containing layer. Small diameter channels restricted the Reynolds
number of fluids to cause laminar flow, and the size of droplets at each junction was controlled by the relative
flow rates of solutions at the flow focusing junction (Fig 4E, 4F).

Experimentally, fluid flow was controlled by three independent pressure-driven pumps which
interface with the PDMS device through small-diameter solvent-resistant tubing (Fig 4G)[54]. In agreement
with computational predictions, droplet size was controlled by the relative pressures of the interacting phases.
At a constant discrete phase pressure, as the continuous phase flow pressure was increased droplet size was
decreased (Fig 4H). Droplet formation in the device was visualized by light microscope, demonstrating the
rapid formation of discrete droplets at each flow-focusing junction (Fig 4I, Sup Movie 1). At Junction 1 (J1),
OS-EXOs suspended in PBS were discretized in a matrix of PLGA-PEG-PLGA, an amphiphilic triblock
copolymer with distinct hydrophilic and hydrophobic blocks, in ethyl acetate (2% wt/v) to form a series of
water-organic (w/o) droplets (Fig 4J). At Junction 2 (J2), the w/o droplet becomes discretized in dilute
polyvinyl alcohol (PVA, 0.01% wt/v), a non-ionic surfactant which stabilized the organic-aqueous interface
during droplet formation in the jet of newly formed w/o/w droplets (Fig 4K).

Droplets were collected from the device in a round-bottom flask and ethyl acetate was evaporated to
yield solid polymeric spheres which were uniform in size, with smooth, non-porous surface morphology (Fig
5A). EXO-MS were fabricated with fluorescently-labelled copolymer blocks, as described in the Methods
section, and visualized by confocal microscopy (Fig 5B). The particles consisted largely of a disperse PLGA
matrix (red), the main component of the triblock PLGA-PEG-PLGA copolymer. An inner layer of PEG
formed a thin spherical layer (blue), wherein OS-EXOs were localized in the particle (green). EXO-MS

18
fabricated by droplet microfluidics could yield smaller and more uniformly sized particles than those
fabricated by a standard mechanical emulsion technique (Fig 5C&D).

3.4. Three-Dimensional Exosome-Releasing Tissue Engineering Constructs

EXO-MS are sufficiently small such that they can be physically attached to the surface of a tissue
engineering scaffold, as shown schematically in Fig 5E. Three-dimensional scaffolds were fabricated from
poly(L-lactic acid) (PLLA) (Fig 5F), which have a nanofibrous surface texture to mimic the native collagen
extracellular matrix with an average fiber diameter of 50-500 nm, the result of thermally induced phase
separation of PLLA [28]. In this work, the scaffolds were fabricated with 250-425 µm sized spherical
macropores using a sugar sphere templating technique [27].

The high surface area of interconnected macropores is ideal to be functionalized with inductive cues
which guide tissue formation in an efficient way. EXO-MS were attached to the surface of scaffold pores using
a post-fabrication seeding method (Fig 5G) [24]. Observed by light microscopy, particles (colored blue for
visualization) were distributed evenly throughout the pores of the matrix. EXO-MS were sufficiently small
such that they did not occlude interconnections between pores which would otherwise limit mass transport and
cell migration throughout the construct (Fig 5H&I).

3.5. Characterizing the Release of OS-EXOs from EXO-MS

OS-EXO encapsulating polymeric microspheres (EXO-MS) were fabricated with four different
formulations of triblock copolymer. PLGA-b-PEG-b-PLGA was synthesized at lactide/glycolide percentages
of 100/0, 85/15, 70/30, and 50/50, all with the same molecular weight (5 kDa). EXO-MS were incubated in
PBS and aliquots analyzed for OS-EXOs. Cumulative release from EXO-MS in PBS showed an initial burst
release followed by a nearly linear release profile, delivering a consistent dose of exosomes for up to 10 weeks
(Fig 6A). Within the first day 18% was released from PL85G15A EXO-MS; following this initial burst release,
the release rate became nearly constant, which is correlated to the hydrophilicity of the PLGA copolymer
segments. The burst release was more prominent in more hydrophilic compositions compared to more
hydrophobic compositions suggesting that they were more easily hydrolyzed.

EXO-MS attached to PLLA scaffolds had a dampened release profile compared to EXO-MS
suspended freely in solution (Fig 6B). The release profiles showed a significantly decreased initial burst
release (7% release in 24 hours) and more linear, consistent controlled release throughout the duration, again
correlated with relative hydrophilicity of the PLGA copolymer segments. Compared to free EXO-MS, we
observed that EXO-MS attached to scaffolds released for a longer period of time despite the same initial cargo

19
loading and polymer composition. PL85G15A composition was chosen for subsequent in vitro and in vivo study
because it maximizes the nearly linear delivery.

3.6. OS-EXO Integrity following Encapsulation and Release

At designated time points, the conditioned OS-EXO-containing PBS was withdrawn and frozen at -
80℃ for storage, and replaced with fresh PBS before analysis. Released OS-EXO concentration was assayed
using NTA (nanotracking analysis). OS-EXOs maintained their characteristic hydrodynamic diameter
throughout release (Fig 6C, Sup Fig 4). Fluorescently-labelled OS-EXOs, eluted from EXO-MS, were readily
up taken by recipient mBMSCs within 30 minutes (Fig 6D). To visualize membrane integrity, eluted exosomes
were fixed and stained for transmission electron microscopy. Compared to virgin OS-EXOs, released OS-
EXOs maintained their spherical shape, size, and membrane integrity, attributed to the addition of D-trehalose
to the discrete exosome-containing aqueous phase. Transmission electron microscopy analysis confirmed that
exosomes maintained their morphology throughout release (Fig 6E). A stabilized lipid membrane was
important to maintain the integrity of the biologic cargo.

miRNA and nucleotide cargo are critical to the physiologic function of exosomes in vitro and in vivo.
Extracted miRNA from various time points was assayed for concentration. The amount of RNA extracted from
eluted OS-EXOs at each release time point matched well to OS-EXO particle number, from NTA, for free
EXO-MS and EXO-MS attached to nanofibrous scaffolds (Fig 6F-I).

3.7. In vitro Bioactivity of Functional EXO-MS Eluent

In vitro bioactivity was investigated over the course of three weeks. mBMSCs were cultured on
nanofibrous scaffolds functionalized with EXO-MS but no exogenous OS-EXO treatment, blank nanofibrous
scaffolds plus exogenous OS-EXO treatments of 1 µg/mL or 5 µg/mL, and blank nanofibrous scaffolds
without any exosome treatment, in osteogenic and growth media. Fourier-transformed infrared spectroscopy
(FTIR) confirmed the presence of calcium hydroxyapatite mineral (Ca10(PO4)6(OH)2) based on characteristic
peaks in the spectra [55]. Constructs exposed to EXO-MS or exogenous OS-EXOs showed strong reflectance
of PO43- (1,185-900 cm-1, 1,035-1,025 cm-1), compared to the positive and negative controls (Fig 7A, 7F).
Carbonate can substitute for OH moieties in apatite (877 cm-1) or phosphate (871 cm-1). Its bands appeared in
the treated constructs, much more intensely than in the negative control. The combination of these signals was
indicative of hydroxyapatite mineralization, which was confirmed by colorimetric calcium assay at 14 and 21
days of culture (Fig 7B, 7G). At 14 days, the EXO-MS-functionalized construct showed significantly greater
mineralization than exogenous administration, which together were greater than the positive and negative

20
control (Fig 7B). By 21 days, the EXO-MS-functionalized constructs were mineralized similarly to the
constructs receiving exogenous treatment; the OS-EXO treated groups were mineralized significantly more
than the positive and negative controls (Fig 7G).

Energy dispersive X-ray spectroscopy (EDX) was used to quantitatively determine the spatial
distribution of elements in the constructs in tissue culture [56]. Compared to the untreated controls, EXO-MS-
functionalized constructs and constructs receiving exogenous OS-EXO treatment showed increased calcium
content (Fig 7C, 7H), and significantly decreased C/P (Fig 7D, 7I) and C/Ca (Fig 7E, 7J) ratios at 14 and 21
days, indicating that the calcium phosphate (CaP) content increased, and the proportion of organic components
decreased as the mineral was deposited.

3.8. Subcutaneous Implantation of EXO-MS-functionalized Scaffolds

Constructs were seeded with primary mBMSCs and implanted subcutaneously in mice for 8 weeks.
Histological analysis by Hematoxylin and Eosin staining indicated that EXO-MS functionalized scaffolds
supported cells similarly to blank nanofibrous scaffolds (Fig 8, Sup. Fig. 5). Masson’s Trichrome staining
showed increased ECM deposition in OS-EXO-loaded EXO-MS (OS-EXO-MS) constructs compared to blank
EXO-MS constructs or blank scaffolds, attributing this difference to the controlled release of OS-EXO cargo.
Here, blank EXO-MS were polymer microspheres loaded with PBS/trehalose without exosomes to serve as a
control for EXO-MS, which were polymer microspheres loaded with exosomes in PBS/trehalose. Von Kossa
staining was more intense in OS-EXO-MS constructs, indicating early mineralization caused by OS-EXO, not
present in blank EXO-MS constructs nor blank scaffolds. The controlled release of OS-EXO also caused
upregulation of bone sialoprotein as assessed by immunohistochemistry.

3.9. Mouse Calvarial Bone Defect Repair by EXO-MS Loaded Scaffolds

OS-EXO-MS-functionalized scaffolds were assessed as a method of cell-free bone regeneration in

vivo in a calvarial defect model. A 2.7-mm round defect was created in the mouse skull. OS-EXO-MS-
functionalized scaffold was used to fill the defect; blank EXO-MS scaffold, and blank scaffold were used as
controls. H&E and trichrome staining showed that the OS-EXO-MS functionalized scaffold contained cell-
laden collagen-rich matrix, filled the defect with significant marrow-containing bone tissue, and was integrated
with the host tissue at 8 weeks (Fig 9A). Untreated controls did not show this integration; while detectable
cells migrated into the construct, they did not form robust osseous tissue. Additionally, no signs of
inflammation were noted, indicating that the constructs, made of biodegradable and biocompatible polymers
which degrade over time, were safe and did not have harmful side effects. Three-dimensional micro-computed

21
tomography (µCT) reconstructions of the skulls showed that localized OS-EXO delivery by EXO-MS resulted
in the best regenerative outcome among treatment groups (Fig 9B). OS-EXO-MS-functionalized scaffold
group showed significant mineralized bone volume in the defect site (Fig 9C), and bone mineral density in the
mineralized areas was similar between the groups (Fig 9D).

4. Discussion

Traditional bone tissue engineering strategies have historically relied on the exogenous transplantation
of stem/progenitor cells with scaffold into a defect site to catalyze tissue neogenesis. Despite the significant
progress achieved, there is limited evidence to support the long-term survival and differentiation of
transplanted stem/progenitor cells in tissue engineering constructs [57]. It is believed that transplanted stem
cells have a strong paracrine effect even if they don’t survive transplantation. Some cell-secreted components
are likely involved in signaling cascades which direct regeneration. Additionally, there are well-known
immunogenic concerns over donor cells, particularly in geriatric and pediatric patients, associated with stem
cell therapies [58, 59]. Therefore, cell-free methods of bone regeneration which recapitulate the biologic
properties of stem/progenitor cells to initiate healing and tissue regeneration are advantageous [24, 39, 45, 52,
60]. Here, we investigated the potential of exosomes derived from mineralizing cells as a new way to induce
bone regeneration in vivo, without requiring the transplantation of exogenous cells. There have been efforts to
utilize EXOs derived from various cell sources either alone or in combination with scaffolds to enhance bone
and other tissue regeneration. For example, EXOs from iPSC-derived MSCs have been combined with
tricalcium phosphate to improve bone regeneration [40]. EXOs derived from gingival or adipose-derived
MSCs have been incorporated in polymer or polymer/calcium silicates composite scaffolds to enhance bone
regeneration [61, 62]. EXOs from different sources have also been incorporated into hydrogels or hydrogel
sponges for bone regeneration [63, 64]. While positive effect of EXO/EVs have been demonstrated, controlling
the release duration and kinetics of EXO remains a challenge [65]. Given that the kinetics of exosome delivery
impacts tissue regeneration [66], we engineered an implantable controlled release platform for their sustained
delivery to maximize their therapeutic benefit in this work.

In order to realize the biologic capabilities of cell-derived exosomes to facilitate tissue regeneration in
the clinical setting, a platform which allows for their controlled delivery and maintains their biologic activity is
needed. We report that: i) human DPSCs release exosomes which are readily isolated, regulating phenotypic
changes in recipient BMSCs by upregulating Erk phosphorylation and enhancing mineralization in both mouse
and human BMSCs; ii) exosomes can be encapsulated in a biodegradable polymer vehicle, EXO-MS, by

22
utilizing block copolymer self-assembly, in a way that maintains their cargo and biologic potency; iii)
exosomes are released from the polymeric microspheres (EXO-MS) in a controlled fashion, and the kinetics of
their release is tunable by varying the polymer composition; and iv) the controlled delivery of OS-EXO via
EXO-MS in vivo facilitates calvarial bone regeneration by recruiting endogenous cells near the defect site and
direct their differentiation and mineralization to form new bone tissue. We believe that this biomaterial
technology, which utilizes the exosomes rather the source stem cells, may have significant translational and
clinical advantages over traditional cell-based regenerative therapies.

A three-dimensional matrix that can support cell infiltration and vascularization is critical to support
tissue neogenesis [39, 67]. Our scaffolds have good mechanical properties, nanofibrous surface texture, and
well interconnected macropores which are shown to be osteogenic in vitro and in vivo, supporting cellular
proliferation and protein adsorption [68-70]. The highly porous surfaces throughout these constructs make
them an ideal candidate to be functionalized with delivery moieties, such as EXO-MS, which provide inductive
cues for regeneration. OS-EXOs are released from the three-dimensional tissue engineering constructs with a
well-controlled rate and duration; the amount of OS-EXOs delivered is modulated by the amount of EXO-MS
incorporated into the scaffold. Additionally, mesenchymal stem/progenitor cell-derived exosomes are known
to be pro-migratory [71]. The early release of OS-EXOs likely attracts endogenous cells from nearby host
tissue to the defect site to secret an early extracellular matrix within the construct which helps orchestrate the
formation of new bone tissue.

Microfluidic droplet generation and polymeric self-assembly are critical to the controlled properties of
EXO-MS, both in their size and facilitating controlled release of OS-EXOs. The small and uniform size of
EXO-MS is advantageous to subtly attach the delivery component to the scaffold pore walls. Microfluidic
droplet formation driven by periodic accretion of pressure forms a free droplet from the dispersed phase within
a continuous phase. Under restricted fluid flow surface tension is dominant over viscous forces, and small
droplets are formed in the device channels [72]. We have also demonstrated that highly concentrated
surfactants used in mechanical emulsions are not necessary in microfluidic emulsions.

A core challenge to the delivery of complex biomolecules is maintaining their bioactivity throughout
encapsulation and release [73]. The hydrophilic and hydrophobic blocks of the triblock copolymer used to
fabricate EXO-MS can associate with the aqueous and organic phases, respectively, to form a core-shell
microsphere which creates an exosome compartment on its interior. In this way they act like a surfactant.
Harsh mechanical techniques such as probe sonication, commonly used in double emulsion methods, are
destructive to exosomes [74]. The PEG-block of the copolymer associates with the OS-EXO containing

23
aqueous phase to facilitate its compartmentalization and discretization from the bulk phase. Ethyl acetate is
chosen because of its low interfacial surface tension with water (6.8 x 10-3 N/m). The polymer also likely
protects exosomes from oxidizing agents and enzymes in vivo. Combined with dilute D-trehalose stabilizer (2
mM), the physical and molecular integrity of OS-EXOs is preserved. Finally, our results indicate that
degradation products of PLGA do not significantly impact OS-EXOs stability. In this way, OS-EXOs are
sufficiently preserved and able to modulate downstream activity of cells in the same fashion as exogenous
administration in vitro, through a single platform which is easily implanted in vivo and allows for local,
controlled delivery.

Microfluidics also offers a high degree of batch-to-batch consistency, compared to other mechanical
emulsion techniques. Emulsions typically result in a wide distribution of droplet sizes [75]. Particle size and
more specifically surface area to volume ratio affect the release profile [76]. A small, immediate burst provides
an immediate therapeutic dose to the defect site, likely influencing migration and differentiation of early
migrant cells into the construct. The following prolonged, nearly zero-order release provides constant exposure
which promotes gradual healing and bone tissue neogenesis, in a similar way to exogenous administration.
Cells are shown to be saturable by exosomes [77]. Therefore, a large burst release is wasteful, and this notion
of saturation provides rationale to maximize the sustained release. Since PLGA is rather hydrophobic, large
exosomes cannot diffuse through the matrix as easily as small molecules or even macromolecules. Therefore
OS-EXOs are released as the polymer sphere is hydrolyzed and becomes porous on a larger scale, maximizing
the linear zero-order release for the majority of the EXO-MS’s functional lifespan.

We have demonstrated that exosomes from mineralizing hDPSCs (OS-EXOs) could enhance MSC
osteogenic differentiation in vitro and bone repair in vivo. OS-EXOs induce the osteogenic activity of BMSCs
as early as 3 days (increased ALP activity) and into late stages (increased mineralization output) in both mouse
and human cells. Consistently, our in vivo models demonstrated the OS-EXOs released from scaffold share the
same bioactivity by accelerating the secretion of extracellular matrix in subcutaneously implanted constructs,
and craniofacial bone repair in a calvarial defect.

In recent years, exosomes derived from mesenchymal stem cells have shown great promise in accelerating
musculoskeletal tissue healing. Most of these reports focus on exosomes derived from naïve MSCs which
facilitate tissue repair by enhancing cell migration, survival or proliferation in a dose-dependent manner, but
require relatively high concentrations in vitro (10-100 µg/mL) [17, 78]. Enhanced regeneration is likely to be
achieved by greater cell infiltration. The application of naïve exosomes may be limited by high cost and
unknown side-effects associated with high doses required in vivo. Here we report a low therapeutic dose (1-5

24
µg/mL) of OS-EXOs which specifically enhances new bone regeneration by inducing osteogenic
differentiation. Additionally, because hDPSCs are isolated from extracted teeth, they are an easily accessible
mesenchymal stromal cell source compared to bone marrow or adipose tissue which require invasive biopsies.

Compared with naïve exosomes, the cargo in exosomes derived from mature cells changes as a function of
differentiation [79, 80]. The osteogenic-specific gene expression and protein synthesis in EXO-MS treated
cells is superior compared to treatment by osteogenic media alone. Runx2 is a key osteogenic differentiation
regulator previously shown to be implicated in exosome-cell interactions, and enhance the transcriptional
activity and stabilization of Runx2 in osteogenic differentiation [81-83]. We demonstrated that with OS-EXO
treatment, Runx2 was upregulated with a synchronous increase in Erk phosphorylation. The Erk pathway is
known to be involved in cell-exosome interactions as well as the activation and stabilization of Runx2 [83, 84].
Interestingly, the BMP related Smad1 pathway is not involved. Since Runx2 coordinates multiple pathways
which may incur complex crosstalk during osteogenic differentiation, a more detailed mechanism of the
specific signaling pathways involved in OS-EXO-mediated osteogenesis merits further investigation to
identify OS-EXO-specific miRNA cargo. We have additionally demonstrated that exosomes derived from
human cells are able to modulate the phenotype of recipient cells similarly in both humans and mice, which
agrees with recent literature indicating cross-species efficacy of exosomes [85].

Our ability to encapsulate and deliver stem cell-derived exosomes demonstrates clear translational
potential towards exciting clinical developments in regenerative medicine. Donor cell phenotype is carefully
selected to direct recipient cells on a designated trajectory, through their exosomal miRNA cargo. Following
this rationale, we are able to hack into endogenous repair mechanisms and potentially circumvent costly stem
cell transplantation therapies. In support of this notion, exosome-mediated cell-free therapy has a number of
manufacturing and regulatory advantages compared to stem cell therapies [19, 86, 87]. Of importance, EXO-
MS and scaffold constructs can be conveniently sterilized by ethylene oxide gas.

5. Conclusions

Exosomes derived from hDPSCs are encapsulated in triblock PLGA-PEG-PLGA microspheres utilizing a
microfluidic device, which are subsequently incorporated into a nanofibrous PLLA scaffold. The controlled
delivery of the exosomes from the scaffold in a critical-size bone defect recruits endogenous cells and
stimulates bone tissue neogenesis in vivo, without the need for exogenous stem cell transplantation. This
platform facilitates the local activity of exosomes, which bear many of the favorable properties of stem cells,

25
without off-target effects or immunogenic and tumorigenic concerns. Exosomes could facilitate a new wave of
potential therapies in regenerative medicine; their clinical success will be highly dependent on methods
developed for their efficient delivery. Our results provide strong evidence to suggest that exosomes derived
from mineralizing hDPSCs provide pro-mineralization cues to drive local stem/progenitor cells towards
osteogenic differentiation in vitro and in vivo.

Supplementary Information is attached.

Acknowledgments

The work was carried out in the Polymeric Biomaterials and Tissue Engineering Laboratory at the University
of Michigan. The authors thank Tianming Du and Youlong Xie for their assistance and discussion in TEM
analysis and cell culture studies. The authors would like to acknowledge the financial support from the NIH
R01-DE022327 (PXM), R01-DE27662 (PXM), T32 DE-007057 (WBS). ZZ was partially supported by China
Scholarship Council (CSC201706240064) for living expenses.

Disclosures

The authors have no competing financial or other interests to disclose related to the present work.

CRediT author statement

This work was carried out in the Polymeric Biomaterials and Tissue Engineering Laboratory at the University
of Michigan. WBS and ZZ contributed overall equally to this work and drafted the manuscript. ZZ did mainly
the biological and animal experiments including exosome isolation, cell culture, molecular and biochemical
characterizations, subcutaneous implantation, calvarial bone defect repair animal surgeries, and histological
analysis. WBS did mainly the biomaterials work, including polymer synthesis, scaffold and microsphere
fabrication, controlled release, and physical characterizations. KX designed and fabricated the microfluidics
devices. TG participated and contributed significantly in animal and biological studies. ME and ZW
participated in biomaterials and controlled release experiments. PXM organized and supervised the project,
trained study participants, obtained funding, analyzed and interpreted data with trainees, revised and finalized
the manuscript, and is the corresponding author.

26
Declaration of interests

☒ The authors declare that they have no known competing financial interests or personal relationships that could
have appeared to influence the work reported in this paper.

References
[1] J.A. Buza, 3rd, T. Einhorn, Bone healing in 2016, Clin Cases Miner Bone Metab 13(2)
(2016) 101-105.
[2] A. Nauth, M.D. McKee, T.A. Einhorn, J.T. Watson, R. Li, E.H. Schemitsch, Managing bone
defects, J Orthop Trauma 25(8) (2011) 462-6.
[3] R. Langer, J.P. Vacanti, Tissue engineering, Science 260(5110) (1993) 920-6.
[4] P.X. Ma, Biomimetic materials for tissue engineering, Adv Drug Deliv Rev 60(2) (2008)
184-198.
[5] C.T. Laurencin, A.M. Ambrosio, M.D. Borden, J.A. Cooper, Jr., Tissue engineering:
orthopedic applications, Annu Rev Biomed Eng 1 (1999) 19-46.
[6] B. Baroli, From natural bone grafts to tissue engineering therapeutics: Brainstorming on
pharmaceutical formulative requirements and challenges, J Pharm Sci 98(4) (2009) 1317-75.
[7] I.J. De Kok, S.J. Peter, M. Archambault, C. van den Bos, S. Kadiyala, I. Aukhil, L.F. Cooper,
Investigation of allogeneic mesenchymal stem cell-based alveolar bone formation:
preliminary findings, Clin Oral Implants Res 14(4) (2003) 481-9.
[8] T.R. Heathman, A.W. Nienow, M.J. McCall, K. Coopman, B. Kara, C.J. Hewitt, The
translation of cell-based therapies: clinical landscape and manufacturing challenges, Regen
Med 10(1) (2015) 49-64.
[9] C. Thery, K.W. Witwer, E. Aikawa, M.J. Alcaraz, J.D. Anderson, R. Andriantsitohaina, A.
Antoniou, T. Arab, F. Archer, G.K. Atkin-Smith, D.C. Ayre, J.M. Bach, D. Bachurski, H.
Baharvand, L. Balaj, S. Baldacchino, N.N. Bauer, A.A. Baxter, M. Bebawy, C. Beckham, A.
Bedina Zavec, A. Benmoussa, A.C. Berardi, P. Bergese, E. Bielska, C. Blenkiron, S. Bobis-
Wozowicz, E. Boilard, W. Boireau, A. Bongiovanni, F.E. Borras, S. Bosch, C.M. Boulanger,
X. Breakefield, A.M. Breglio, M.A. Brennan, D.R. Brigstock, A. Brisson, M.L. Broekman,
J.F. Bromberg, P. Bryl-Gorecka, S. Buch, A.H. Buck, D. Burger, S. Busatto, D. Buschmann,
B. Bussolati, E.I. Buzas, J.B. Byrd, G. Camussi, D.R. Carter, S. Caruso, L.W. Chamley, Y.T.
Chang, C. Chen, S. Chen, L. Cheng, A.R. Chin, A. Clayton, S.P. Clerici, A. Cocks, E.
Cocucci, R.J. Coffey, A. Cordeiro-da-Silva, Y. Couch, F.A. Coumans, B. Coyle, R.
Crescitelli, M.F. Criado, C. D'Souza-Schorey, S. Das, A. Datta Chaudhuri, P. de Candia, E.F.
De Santana, O. De Wever, H.A. Del Portillo, T. Demaret, S. Deville, A. Devitt, B. Dhondt,
D. Di Vizio, L.C. Dieterich, V. Dolo, A.P. Dominguez Rubio, M. Dominici, M.R. Dourado,
T.A. Driedonks, F.V. Duarte, H.M. Duncan, R.M. Eichenberger, K. Ekstrom, S. El
Andaloussi, C. Elie-Caille, U. Erdbrugger, J.M. Falcon-Perez, F. Fatima, J.E. Fish, M.
Flores-Bellver, A. Forsonits, A. Frelet-Barrand, F. Fricke, G. Fuhrmann, S. Gabrielsson, A.
Gamez-Valero, C. Gardiner, K. Gartner, R. Gaudin, Y.S. Gho, B. Giebel, C. Gilbert, M.
Gimona, I. Giusti, D.C. Goberdhan, A. Gorgens, S.M. Gorski, D.W. Greening, J.C. Gross, A.
Gualerzi, G.N. Gupta, D. Gustafson, A. Handberg, R.A. Haraszti, P. Harrison, H. Hegyesi,
A. Hendrix, A.F. Hill, F.H. Hochberg, K.F. Hoffmann, B. Holder, H. Holthofer, B.
Hosseinkhani, G. Hu, Y. Huang, V. Huber, S. Hunt, A.G. Ibrahim, T. Ikezu, J.M. Inal, M.

27
 Isin, A. Ivanova, H.K. Jackson, S. Jacobsen, S.M. Jay, M. Jayachandran, G. Jenster, L. Jiang,
S.M. Johnson, J.C. Jones, A. Jong, T. Jovanovic-Talisman, S. Jung, R. Kalluri, S.I. Kano, S.
Kaur, Y. Kawamura, E.T. Keller, D. Khamari, E. Khomyakova, A. Khvorova, P. Kierulf,
K.P. Kim, T. Kislinger, M. Klingeborn, D.J. Klinke, 2nd, M. Kornek, M.M. Kosanovic, A.F.
Kovacs, E.M. Kramer-Albers, S. Krasemann, M. Krause, I.V. Kurochkin, G.D. Kusuma, S.
Kuypers, S. Laitinen, S.M. Langevin, L.R. Languino, J. Lannigan, C. Lasser, L.C. Laurent,
G. Lavieu, E. Lazaro-Ibanez, S. Le Lay, M.S. Lee, Y.X.F. Lee, D.S. Lemos, M. Lenassi, A.
Leszczynska, I.T. Li, K. Liao, S.F. Libregts, E. Ligeti, R. Lim, S.K. Lim, A. Line, K.
Linnemannstons, A. Llorente, C.A. Lombard, M.J. Lorenowicz, A.M. Lorincz, J. Lotvall, J.
Lovett, M.C. Lowry, X. Loyer, Q. Lu, B. Lukomska, T.R. Lunavat, S.L. Maas, H. Malhi, A.
Marcilla, J. Mariani, J. Mariscal, E.S. Martens-Uzunova, L. Martin-Jaular, M.C. Martinez,
V.R. Martins, M. Mathieu, S. Mathivanan, M. Maugeri, L.K. McGinnis, M.J. McVey, D.G.
Meckes, Jr., K.L. Meehan, I. Mertens, V.R. Minciacchi, A. Moller, M. Moller Jorgensen, A.
Morales-Kastresana, J. Morhayim, F. Mullier, M. Muraca, L. Musante, V. Mussack, D.C.
Muth, K.H. Myburgh, T. Najrana, M. Nawaz, I. Nazarenko, P. Nejsum, C. Neri, T. Neri, R.
Nieuwland, L. Nimrichter, J.P. Nolan, E.N. Nolte-'t Hoen, N. Noren Hooten, L. O'Driscoll,
T. O'Grady, A. O'Loghlen, T. Ochiya, M. Olivier, A. Ortiz, L.A. Ortiz, X. Osteikoetxea, O.
Ostergaard, M. Ostrowski, J. Park, D.M. Pegtel, H. Peinado, F. Perut, M.W. Pfaffl, D.G.
Phinney, B.C. Pieters, R.C. Pink, D.S. Pisetsky, E. Pogge von Strandmann, I. Polakovicova,
I.K. Poon, B.H. Powell, I. Prada, L. Pulliam, P. Quesenberry, A. Radeghieri, R.L. Raffai, S.
Raimondo, J. Rak, M.I. Ramirez, G. Raposo, M.S. Rayyan, N. Regev-Rudzki, F.L. Ricklefs,
P.D. Robbins, D.D. Roberts, S.C. Rodrigues, E. Rohde, S. Rome, K.M. Rouschop, A.
Rughetti, A.E. Russell, P. Saa, S. Sahoo, E. Salas-Huenuleo, C. Sanchez, J.A. Saugstad, M.J.
Saul, R.M. Schiffelers, R. Schneider, T.H. Schoyen, A. Scott, E. Shahaj, S. Sharma, O.
Shatnyeva, F. Shekari, G.V. Shelke, A.K. Shetty, K. Shiba, P.R. Siljander, A.M. Silva, A.
Skowronek, O.L. Snyder, 2nd, R.P. Soares, B.W. Sodar, C. Soekmadji, J. Sotillo, P.D. Stahl,
W. Stoorvogel, S.L. Stott, E.F. Strasser, S. Swift, H. Tahara, M. Tewari, K. Timms, S.
Tiwari, R. Tixeira, M. Tkach, W.S. Toh, R. Tomasini, A.C. Torrecilhas, J.P. Tosar, V.
Toxavidis, L. Urbanelli, P. Vader, B.W. van Balkom, S.G. van der Grein, J. Van Deun, M.J.
van Herwijnen, K. Van Keuren-Jensen, G. van Niel, M.E. van Royen, A.J. van Wijnen, M.H.
Vasconcelos, I.J. Vechetti, Jr., T.D. Veit, L.J. Vella, E. Velot, F.J. Verweij, B. Vestad, J.L.
Vinas, T. Visnovitz, K.V. Vukman, J. Wahlgren, D.C. Watson, M.H. Wauben, A. Weaver,
J.P. Webber, V. Weber, A.M. Wehman, D.J. Weiss, J.A. Welsh, S. Wendt, A.M. Wheelock,
Z. Wiener, L. Witte, J. Wolfram, A. Xagorari, P. Xander, J. Xu, X. Yan, M. Yanez-Mo, H.
Yin, Y. Yuana, V. Zappulli, J. Zarubova, V. Zekas, J.Y. Zhang, Z. Zhao, L. Zheng, A.R.
Zheutlin, A.M. Zickler, P. Zimmermann, A.M. Zivkovic, D. Zocco, E.K. Zuba-Surma,
Minimal information for studies of extracellular vesicles 2018 (MISEV2018): a position
statement of the International Society for Extracellular Vesicles and update of the
MISEV2014 guidelines, J Extracell Vesicles 7(1) (2018) 1535750.
[10] K.W. Witwer, B.W.M. Van Balkom, S. Bruno, A. Choo, M. Dominici, M. Gimona, A.F.
Hill, D. De Kleijn, M. Koh, R.C. Lai, S.A. Mitsialis, L.A. Ortiz, E. Rohde, T. Asada, W.S.
Toh, D.J. Weiss, L. Zheng, B. Giebel, S.K. Lim, Defining mesenchymal stromal cell (MSC)-
derived small extracellular vesicles for therapeutic applications, J Extracell Vesicles 8(1)
(2019) 1609206.
[11] W.B. Swanson, T. Gong, Z. Zhang, M. Eberle, D. Niemann, R. Dong, K.J. Rambhia, P.X.
Ma, Controlled release of odontogenic exosomes from a biodegradable vehicle mediates

28
 dentinogenesis as a novel biomimetic pulp capping therapy, J Control Release 324 (2020)
679-694.
[12] R.M. Johnstone, M. Adam, J.R. Hammond, L. Orr, C. Turbide, Vesicle formation during
reticulocyte maturation. Association of plasma membrane activities with released vesicles
(exosomes), J Biol Chem 262(19) (1987) 9412-20.
[13] M. Tkach, C. Thery, Communication by Extracellular Vesicles: Where We Are and Where
We Need to Go, Cell 164(6) (2016) 1226-1232.
[14] B. Fevrier, G. Raposo, Exosomes: endosomal-derived vesicles shipping extracellular
messages, Curr Opin Cell Biol 16(4) (2004) 415-21.
[15] P.J. Quesenberry, J.M. Aliotta, Cellular phenotype switching and microvesicles, Adv Drug
Deliv Rev 62(12) (2010) 1141-8.
[16] R. Kishore, M. Khan, More Than Tiny Sacks: Stem Cell Exosomes as Cell-Free Modality
for Cardiac Repair, Circ Res 118(2) (2016) 330-43.
[17] S. Zhang, S.J. Chuah, R.C. Lai, J.H.P. Hui, S.K. Lim, W.S. Toh, MSC exosomes mediate
cartilage repair by enhancing proliferation, attenuating apoptosis and modulating immune
reactivity, Biomaterials 156 (2018) 16-27.
[18] P. Sharma, P. Mesci, C. Carromeu, D.R. McClatchy, L. Schiapparelli, J.R. Yates, A.R.
Muotri, H.T. Cline, Exosomes regulate neurogenesis and circuit assembly, Proceedings of the
National Academy of Sciences 116(32) (2019) 16086-16094.
[19] S. Rani, A.E. Ryan, M.D. Griffin, T. Ritter, Mesenchymal Stem Cell-derived Extracellular
Vesicles: Toward Cell-free Therapeutic Applications, Mol Ther 23(5) (2015) 812-823.
[20] C.P. Lai, O. Mardini, M. Ericsson, S. Prabhakar, C.A. Maguire, J.W. Chen, B.A. Tannous,
X.O. Breakefield, Dynamic Biodistribution of Extracellular Vesicles in Vivo Using a
Multimodal Imaging Reporter, ACS Nano 8(1) (2014) 483-494.
[21] R. Langer, Drug delivery and targeting, Nature 392(6679 Suppl) (1998) 5-10.
[22] F. Danhier, E. Ansorena, J.M. Silva, R. Coco, A. Le Breton, V. Préat, PLGA-based
nanoparticles: An overview of biomedical applications, Journal of Controlled Release 161(2)
(2012) 505-522.
[23] G. Wei, G.J. Pettway, L.K. McCauley, P.X. Ma, The release profiles and bioactivity of
parathyroid hormone from poly(lactic-co-glycolic acid) microspheres, Biomaterials 25(2)
(2004) 345-52.
[24] G. Wei, Q. Jin, W.V. Giannobile, P.X. Ma, The enhancement of osteogenesis by nano-
fibrous scaffolds incorporating rhBMP-7 nanospheres, Biomaterials 28(12) (2007) 2087-96.
[25] G. Feng, Z. Zhang, M. Dang, X. Zhang, Y. Doleyres, Y. Song, D. Chen, P.X. Ma, Injectable
nanofibrous spongy microspheres for NR4A1 plasmid DNA transfection to reverse fibrotic
degeneration and support disc regeneration, Biomaterials 131 (2017) 86-97.
[26] K. Feng, H. Sun, M.A. Bradley, E.J. Dupler, W.V. Giannobile, P.X. Ma, Novel antibacterial
nanofibrous PLLA scaffolds, J Control Release 146(3) (2010) 363-9.
[27] G. Wei, P.X. Ma, Macroporous and nanofibrous polymer scaffolds and polymer/bone-like
apatite composite scaffolds generated by sugar spheres, J Biomed Mater Res A 78(2) (2006)
306-15.
[28] P.X. Ma, R. Zhang, Synthetic nano-scale fibrous extracellular matrix, J Biomed Mater Res
46(1) (1999) 60-72.
[29] Q. Jin, P.X. Ma, W.V. Giannobile, Platelet-Derived Growth Factor Delivery via
Nanofibrous Scaffolds for Soft-Tissue Repair, Adv Skin Wound Care 1 (2010) 375-381.

29
[30] R. Kuang, Z. Zhang, X. Jin, J. Hu, M.J. Gupte, L. Ni, P.X. Ma, Nanofibrous spongy
microspheres enhance odontogenic differentiation of human dental pulp stem cells, Adv
Healthc Mater 4(13) (2015) 1993-2000.
[31] R. Kuang, Z. Zhang, X. Jin, J. Hu, S. Shi, L. Ni, P.X. Ma, Nanofibrous spongy microspheres
for the delivery of hypoxia-primed human dental pulp stem cells to regenerate vascularized
dental pulp, Acta Biomater 33 (2016) 225-34.
[32] J. Wang, X. Liu, X. Jin, H. Ma, J. Hu, L. Ni, P.X. Ma, The odontogenic differentiation of
human dental pulp stem cells on nanofibrous poly(L-lactic acid) scaffolds in vitro and in
vivo, Acta Biomaterialia 6 (2010) 3856-3863.
[33] W. Wang, M. Dang, Z. Zhang, J. Hu, T.W. Eyster, L. Ni, P.X. Ma, Dentin regeneration by
stem cells of apical papilla on injectable nanofibrous microspheres and stimulated by
controlled BMP-2 release, Acta Biomater 36 (2016) 63-72.
[34] M. Ryner, K. Stridsberg, A.-C. Albertsson, H. von Schenck, M. Svensson, Mechanism of
Ring-Opening Polymerization of 1,5-Dioxepan-2-one andl-Lactide with Stannous 2-
Ethylhexanoate. A Theoretical Study, Macromolecules 34(12) (2001) 3877-3881.
[35] X. Liu, P.X. Ma, The nanofibrous architecture of poly(L-lactic acid)-based functional
copolymers, Biomaterials 31(2) (2010) 259-69.
[36] A.H. Demond, A.S. Lindner, Estimation of Interfacial-Tension between Organic Liquids
and Water, Environ Sci Technol 27(12) (1993) 2318-2331.
[37] S. Gronthos, M. Mankani, J. Brahim, P. Robey, S. Shi, Postnatal human dental pulp stem
cells (DPSCs) in vitro and in vivo, Proceedings of the National Academy of Sciences of the
United States of America 97(25) (2000) 13625-13630.
[38] J. Wang, H. Ma, X. Jin, J. Hu, X. Liu, L. Ni, P.X. Ma, The effect of scaffold architecture on
odontogenic differentiation of human dental pulp stem cells, Biomaterials 32(31) (2011)
7822-30.
[39] X. Zhang, Y. Li, Y.E. Chen, J. Chen, P.X. Ma, Cell-free 3D scaffold with two-stage
delivery of miRNA-26a to regenerate critical-sized bone defects, Nat Commun 7 (2016)
10376.
[40] J. Zhang, X. Liu, H. Li, C. Chen, B. Hu, X. Niu, Q. Li, B. Zhao, Z. Xie, Y. Wang,
Exosomes/tricalcium phosphate combination scaffolds can enhance bone regeneration by
activating the PI3K/Akt signaling pathway, Stem Cell Res Ther 7(1) (2016) 136.
[41] K.O. Jung, H. Jo, J.H. Yu, S.S. Gambhir, G. Pratx, Development and MPI tracking of novel
hypoxia-targeted theranostic exosomes, Biomaterials 177 (2018) 139-148.
[42] D.C. Rio, M. Ares, Jr., G.J. Hannon, T.W. Nilsen, Purification of RNA using TRIzol (TRI
reagent), Cold Spring Harb Protoc 2010(6) (2010) pdb prot5439.
[43] K.J. Livak, T.D. Schmittgen, Analysis of relative gene expression data using real-time
quantitative PCR and the 2(-Delta Delta C(T)) Method, Methods 25(4) (2001) 402-8.
[44] T. Mahmood, P.C. Yang, Western blot: technique, theory, and trouble shooting, N Am J
Med Sci 4(9) (2012) 429-34.
[45] G. Wei, Q. Jin, W.V. Giannobile, P.X. Ma, Nano-fibrous scaffold for controlled delivery of
recombinant human PDGF-BB, J Control Release 112(1) (2006) 103-10.
[46] C. Thery, Isolation and Characterization of Exosomes from Cell Culture Supernatants and
Biological Fluids, Current Protocols in Cell Biology, John Wiley & Sons, Inc.2006, pp.
3.22.1-3.22.29.

30
[47] J. Asadi, S. Ferguson, H. Raja, C. Hacker, P. Marius, R. Ward, C. Pliotas, J. Naismith, J.
Lucocq, Enhanced imaging of lipid rich nanoparticles embedded in methylcellulose films for
transmission electron microscopy using mixtures of heavy metals, Micron 99 (2017) 40-48.
[48] Y.T. Tang, Y.Y. Huang, L. Zheng, S.H. Qin, X.P. Xu, T.X. An, Y. Xu, Y.S. Wu, X.M. Hu,
B.H. Ping, Q. Wang, Comparison of isolation methods of exosomes and exosomal RNA from
cell culture medium and serum, Int J Mol Med 40(3) (2017) 834-844.
[49] N. Nuti, C. Corallo, B.M. Chan, M. Ferrari, B. Gerami-Naini, Multipotent Differentiation of
Human Dental Pulp Stem Cells: a Literature Review, Stem Cell Rev 12(5) (2016) 511-523.
[50] E. van der Pol, A.G. Hoekstra, A. Sturk, C. Otto, T.G. van Leeuwen, R. Nieuwland, Optical
and non-optical methods for detection and characterization of microparticles and exosomes, J
Thromb Haemost 8(12) (2010) 2596-607.
[51] M.F. Favata, K.Y. Horiuchi, E.J. Manos, A.J. Daulerio, D.A. Stradley, W.S. Feeser, D.E.
Van Dyk, W.J. Pitts, R.A. Earl, F. Hobbs, R.A. Copeland, R.L. Magolda, P.A. Scherle, J.M.
Trzaskos, Identification of a novel inhibitor of mitogen-activated protein kinase kinase, J
Biol Chem 273(29) (1998) 18623-32.
[52] M. Dang, A.J. Koh, X. Jin, L.K. McCauley, P.X. Ma, Local pulsatile PTH delivery
regenerates bone defects via enhanced bone remodeling in a cell-free scaffold, Biomaterials
114 (2017) 1-9.
[53] W. Swanson, Gong, T., Zhang, Z., Xiu, K., Eberle, M., Ma, P.X., Controlled Release of
Odontogenic Exosomes from a Biodegradable Vehicle Mediates Dentinogenesis as a Novel
Biomimetic Pulp Capping Therapy, Under Review (2019).
[54] T. Boudot, Nicolas Petit, N., Ritty, S., Kieffer, C., Flow-rate control in a microfluidic
device, 2012.
[55] T. Sonou, M. Ohya, M. Yashiro, A. Masumoto, Y. Nakashima, T. Ito, T. Mima, S. Negi, H.
Kimura-Suda, T. Shigematsu, Mineral Composition of Phosphate-Induced Calcification in a
Rat Aortic Tissue Culture Model, J Atheroscler Thromb 22(11) (2015) 1197-206.
[56] A. Henmi, H. Okata, T. Anada, M. Yoshinari, Y. Mikami, O. Suzuki, Y. Sasano, Bone
matrix calcification during embryonic and postembryonic rat calvarial development assessed
by SEM-EDX spectroscopy, XRD, and FTIR spectroscopy, J Bone Miner Metab 34(1)
(2016) 41-50.
[57] M. Kitami, M. Kaku, J.M. Rocabado, T. Ida, N. Akiba, K. Uoshima, Prolonged Survival of
Transplanted Osteoblastic Cells Does Not Directly Accelerate the Healing of Calvarial Bone
Defects, J Cell Physiol 231(9) (2016) 1974-82.
[58] J. Poulos, The limited application of stem cells in medicine: a review, Stem Cell Res Ther
9(1) (2018) 1.
[59] S. Diederichs, K.M. Shine, R.S. Tuan, The promise and challenges of stem cell-based
therapies for skeletal diseases: stem cell applications in skeletal medicine: potential, cell
sources and characteristics, and challenges of clinical translation, Bioessays 35(3) (2013)
220-30.
[60] M. Dang, L. Saunders, X. Niu, Y. Fan, P.X. Ma, Biomimetic delivery of signals for bone
tissue engineering, Bone Res 6 (2018) 25.
[61] F. Diomede, A. Gugliandolo, P. Cardelli, I. Merciaro, V. Ettorre, T. Traini, R. Bedini, D.
Scionti, A. Bramanti, A. Nanci, S. Caputi, A. Fontana, E. Mazzon, O. Trubiani, Three-
dimensional printed PLA scaffold and human gingival stem cell-derived extracellular
vesicles: a new tool for bone defect repair, Stem Cell Res Ther 9(1) (2018) 104.

31
[62] M.G. Gandolfi, C. Gardin, F. Zamparini, L. Ferroni, M.D. Esposti, G. Parchi, B. Ercan, L.
Manzoli, F. Fava, P. Fabbri, C. Prati, B. Zavan, Mineral-Doped Poly(L-lactide) Acid
Scaffolds Enriched with Exosomes Improve Osteogenic Commitment of Human Adipose-
Derived Mesenchymal Stem Cells, Nanomaterials (Basel) 10(3) (2020).
[63] Q. Shi, Z. Qian, D. Liu, J. Sun, X. Wang, H. Liu, J. Xu, X. Guo, GMSC-Derived Exosomes
Combined with a Chitosan/Silk Hydrogel Sponge Accelerates Wound Healing in a Diabetic
Rat Skin Defect Model, Front Physiol 8 (2017) 904.
[64] K. Zhang, X. Zhao, X. Chen, Y. Wei, W. Du, Y. Wang, L. Liu, W. Zhao, Z. Han, D. Kong,
Q. Zhao, Z. Guo, Z. Han, N. Liu, F. Ma, Z. Li, Enhanced Therapeutic Effects of
Mesenchymal Stem Cell-Derived Exosomes with an Injectable Hydrogel for Hindlimb
Ischemia Treatment, ACS Appl Mater Interfaces 10(36) (2018) 30081-30091.
[65] A. Akbari, N. Jabbari, R. Sharifi, M. Ahmadi, A. Vahhabi, S.J. Seyedzadeh, M. Nawaz, S.
Szafert, M. Mahmoodi, E. Jabbari, R. Asghari, J. Rezaie, Free and hydrogel encapsulated
exosome-based therapies in regenerative medicine, Life Sci 249 (2020) 117447.
[66] H. Henriques-Antunes, R.M.S. Cardoso, A. Zonari, J. Correia, E.C. Leal, A. Jimenez-Balsa,
M.M. Lino, A. Barradas, I. Kostic, C. Gomes, J.M. Karp, E. Carvalho, L. Ferreira, The
Kinetics of Small Extracellular Vesicle Delivery Impacts Skin Tissue Regeneration, ACS
Nano 13(8) (2019) 8694-8707.
[67] M.J. Gupte, W.B. Swanson, J. Hu, X. Jin, H. Ma, Z. Zhang, Z. Liu, K. Feng, G. Feng, G.
Xiao, N. Hatch, Y. Mishina, P.X. Ma, Pore size directs bone marrow stromal cell fate and
tissue regeneration in nanofibrous macroporous scaffolds by mediating vascularization, Acta
Biomater 82 (2018) 1-11.
[68] K.M. Woo, V.J. Chen, P.X. Ma, Nano-fibrous scaffolding architecture selectively enhances
protein adsorption contributing to cell attachment, J Biomed Mater Res A 67(2) (2003) 531-
7.
[69] L.A. Smith, X. Liu, J. Hu, P.X. Ma, The enhancement of human embryonic stem cell
osteogenic differentiation with nano-fibrous scaffolding, Biomaterials 31(21) (2010) 5526-
35.
[70] J. Hu, L.A. Smith, K. Feng, X. Liu, H. Sun, P.X. Ma, Response of human embryonic stem
cell-derived mesenchymal stem cells to osteogenic factors and architectures of materials
during in vitro osteogenesis, Tissue Eng Part A 16(11) (2010) 3507-14.
[71] Y. Zhu, Y. Jia, Y. Wang, J. Xu, Y. Chai, Impaired Bone Regenerative Effect of Exosomes
Derived from Bone Marrow Mesenchymal Stem Cells in Type 1 Diabetes, Stem Cells Transl
Med 8(6) (2019) 593-605.
[72] R. Seemann, M. Brinkmann, T. Pfohl, S. Herminghaus, Droplet based microfluidics, Rep
Prog Phys 75(1) (2012) 016601.
[73] L. Wang, Y. Liu, W. Zhang, X. Chen, T. Yang, G. Ma, Microspheres and microcapsules for
protein delivery: strategies of drug activity retention, Curr Pharm Des 19(35) (2013) 6340-
52.
[74] S. Sukreet, B.V.R.E. Silva, J. Adamec, J. Cui, J. Zempleni, Sonication and Short-term
Incubation Alter the Content of Bovine Milk Exosome Cargos and Exosome Bioavailability
(OR26-08-19), Current Developments in Nutrition 3(Supplement_1) (2019).
[75] T.A. Prileszky, B.A. Ogunnaike, E.M. Furst, Statistics of droplet sizes generated by a
microfluidic device, AIChE Journal 62(8) (2016) 2923-2928.

32
[76] C. Busatto, J. Pesoa, I. Helbling, J. Luna, D. Estenoz, Effect of particle size, polydispersity
and polymer degradation on progesterone release from PLGA microparticles: Experimental
and mathematical modeling, International Journal of Pharmaceutics 536(1) (2018) 360-369.
[77] C.-C. Huang, R. Narayanan, S. Alapati, S. Ravindran, Exosomes as biomimetic tools for
stem cell differentiation: Applications in dental pulp tissue regeneration, Biomaterials 111
(2016) 103-115.
[78] W. Zhang, X. Bai, B. Zhao, Y. Li, Y. Zhang, Z. Li, X. Wang, L. Luo, F. Han, J. Zhang, S.
Han, W. Cai, L. Su, K. Tao, J. Shi, D. Hu, Cell-free therapy based on adipose tissue stem
cell-derived exosomes promotes wound healing via the PI3K/Akt signaling pathway, Exp
Cell Res 370(2) (2018) 333-342.
[79] J.F. Xu, G.H. Yang, X.H. Pan, S.J. Zhang, C. Zhao, B.S. Qiu, H.F. Gu, J.F. Hong, L. Cao,
Y. Chen, B. Xia, Q. Bi, Y.P. Wang, Altered microRNA expression profile in exosomes
during osteogenic differentiation of human bone marrow-derived mesenchymal stem cells,
PLoS One 9(12) (2014) e114627.
[80] X. Wang, O. Omar, F. Vazirisani, P. Thomsen, K. Ekstrom, Mesenchymal stem cell-derived
exosomes have altered microRNA profiles and induce osteogenic differentiation depending
on the stage of differentiation, PLoS One 13(2) (2018) e0193059.
[81] R.T. Franceschi, G. Xiao, Regulation of the osteoblast-specific transcription factor, Runx2:
responsiveness to multiple signal transduction pathways, J Cell Biochem 88(3) (2003) 446-
54.
[82] J.H. Jun, W.J. Yoon, S.B. Seo, K.M. Woo, G.S. Kim, H.M. Ryoo, J.H. Baek, BMP2-
activated Erk/MAP kinase stabilizes Runx2 by increasing p300 levels and histone
acetyltransferase activity, J Biol Chem 285(47) (2010) 36410-9.
[83] G. Xiao, D. Jiang, P. Thomas, M.D. Benson, K. Guan, G. Karsenty, R.T. Franceschi,
MAPK pathways activate and phosphorylate the osteoblast-specific transcription factor,
Cbfa1, J Biol Chem 275(6) (2000) 4453-9.
[84] C. Ge, G. Xiao, D. Jiang, Q. Yang, N.E. Hatch, H. Roca, R.T. Franceschi, Identification and
functional characterization of ERK/MAPK phosphorylation sites in the Runx2 transcription
factor, J Biol Chem 284(47) (2009) 32533-43.
[85] H. Valadi, K. Ekstrom, A. Bossios, M. Sjostrand, J.J. Lee, J.O. Lotvall, Exosome-mediated
transfer of mRNAs and microRNAs is a novel mechanism of genetic exchange between cells,
Nat Cell Biol 9(6) (2007) 654-9.
[86] M. Gimona, K. Pachler, S. Laner-Plamberger, K. Schallmoser, E. Rohde, Manufacturing of
Human Extracellular Vesicle-Based Therapeutics for Clinical Use, Int J Mol Sci 18(6)
(2017).
[87] T. Lener, M. Gimona, L. Aigner, V. Borger, E. Buzas, G. Camussi, N. Chaput, D.
Chatterjee, F.A. Court, H.A. Del Portillo, L. O'Driscoll, S. Fais, J.M. Falcon-Perez, U.
Felderhoff-Mueser, L. Fraile, Y.S. Gho, A. Gorgens, R.C. Gupta, A. Hendrix, D.M.
Hermann, A.F. Hill, F. Hochberg, P.A. Horn, D. de Kleijn, L. Kordelas, B.W. Kramer, E.M.
Kramer-Albers, S. Laner-Plamberger, S. Laitinen, T. Leonardi, M.J. Lorenowicz, S.K. Lim,
J. Lotvall, C.A. Maguire, A. Marcilla, I. Nazarenko, T. Ochiya, T. Patel, S. Pedersen, G.
Pocsfalvi, S. Pluchino, P. Quesenberry, I.G. Reischl, F.J. Rivera, R. Sanzenbacher, K.
Schallmoser, I. Slaper-Cortenbach, D. Strunk, T. Tonn, P. Vader, B.W. van Balkom, M.
Wauben, S.E. Andaloussi, C. Thery, E. Rohde, B. Giebel, Applying extracellular vesicles
based therapeutics in clinical trials - an ISEV position paper, J Extracell Vesicles 4 (2015)
30087.

33
 Figures

Fig. 1. Isolation and characterization of exosomes derived from dental pulp stem cells (DPSCs).
Exosomes are physically characterized by their size using nanotracking analysis, indicating that
freshly isolated exosomes are of the characteristic hydrodynamic diameter (davg = 135 nm, A),
and uniformly sized (B). Molecular profiling by western blot indicates that the isolated
exosomes are highly enriched for characteristic exosome surface markers not abundant in the
total cell isolate (C). Freshly isolated exosomes were fluorescently labelled, as described, and
their uptake by recipient primary mouse bone marrow stromal cells (mBMSCs) after 30 minutes
incubation and visualized by confocal microscopy (D; blue: DAPI, red: F-actin, green:
membrane-fluorescent exosomes; scale: 50 µm).

34
Fig. 2. In vitro characterization of the osteogenic potential of DPSC-derived exosomes (OS-
EXO) in primary mBMSCs. In as few as three days, OS-EXO cause increased alkaline

35
phosphatase (ALP) activity (A, B), and increased calcium deposition at 14 and 21 days, indicated
by Alizarin red staining (C, D), compared to osteogenic or growth media alone. Exogenous OS-
EXO administration in vitro causes significant osteogenesis at the transcriptional (E) and
translational levels (F, G, H). * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001 (n=4 for
Western blotting and all other analyses, representative images presented).

Fig. 3. Osteogenic exosomes (OS-EXO) derived from DPSCs upregulated mBMSCs Runx2
level through ERK signaling. (A,C): OS-EXO upregulated mBMSCs Runx2 and ERK
phosphorylation at 1、5、10 ug/ml concentrations; (B,D): With presence of ERK inhibitor
u1026, the ERK phosphorylation and Runx2 level were both downregulated. N=4 for Western
blotting analysis and representative images presented.

36
Fig. 4. Development of a
dual flow-focusing
junction microfluidic
device to facilitate
exosome encapsulation.
The device is designed to
facilitate the formation of
a water/oil/water double
emulsion (A) via two flow
focusing junctions (B)
where immiscible solvents
contact and form droplets
which later become
particles (C). CAD
drawings of the device are
used to simulate fluid
flow using COMSOL,
demonstrating the
formation of droplets at
both junctions (D, E).
Droplet size is controlled
by the relative flow rates
of two immiscible
solvents when they
contact at the flow
focusing junction (F).
PDMS devices are
interfaced with pressure-
driven pumps and droplet
formation is visualized
under a light microscope
(G, H, I, J, K). Scale =
100 µm.

37
38
Fig 5. Characterization of EXO-MS. Scanning electron micrograph images of EXO-MS prepared
by microfluidic double emulsion method shows smooth, uniformly sized spheres (A). EXO-MS
made from PLGA-PEG-PLGA block copolymer were mixed with fluorescently labelled polymer
blocks and visualized by confocal microscopy; PEG (blue) segments self-assemble on the inside
of the particle to form a sphere wherein OS-EXO (green) localize, among a disperse PLGA (red)
microspheres (B; scale: 15 um). The size of EXO-MS is measured by a laser diffraction method
(C, davg = 7.70 um). Compared to self-assembling particles from the same polymer made under
mechanical emulsion conditions, particles made by a microfluidic platform are significantly
smaller and less disperse (D). EXO-MS are attached to poly(L-lactic acid) (PLLA) tissue
engineering scaffolds, as shown in the schematic (E; scale: 100 um). Scaffolds are a matrix of
interconnected macropores with a nanofibrous surface texture (F). EXO-MS are seeded onto the
pore surface of the scaffold using a solvent-wetting method, visualized by light microscopy (G,
dyed blue for visual contrast) and scanning electron microscopy (H, I).

39
Fig. 6. Evaluation of the
controlled release of OS-EXO
from EXO-MS. The release
profile of OS-EXO is modulated
by the hydrophilicity of the
copolymer (A). After attachment
of EXO-MS to PLLA scaffolds,
the release of OS-EXO becomes
more linear (B). OS-EXO
released from EXO-MS at 2
weeks maintain their
characteristic size (davg = 105
nm), evaluated by nanoparticle
tracking analysis (C). Primary
mBMSCs were treated with
EXO-MS eluent (collected at 2
weeks) which contained
fluorescently labeled OS-EXO
for 30 minutes, and the uptake of
OS-EXO by recipient cells was
visualized by confocal
microscopy (D; green:
fluorescent-membrane of OS-
EXO, red: F-actin, blue: DAPI;
scale: 100 um). Aliquots of in
vitro EXO-MS eluent were
collected and subjected to
transmission electron
microscopy (TEM) analysis at
prescribed time points (E; scale
bars: 100 nm). Aliquots of in
vitro EXO-MS eluent are
subjected to RNA extraction at
each release time point. Quantity
of RNA correlates well to
particle number (using NTA) for
cases of both release from EXO-
MS, and release from EXO-MS
attached to PLLA scaffolds (F,
G, H, I).

40
 41
Fig. 7. Mineral characterization of cell-laden 3D constructs cultured in vitro. At 14 and 21 days of primary mBMSC cell culture,
Fourier-transformed infrared spectroscopy (FTIR) spectra indicate that EXO-MS laden PLLA scaffolds and exogenous administration
of OS-EXO yield calcium hydroxyapatite mineralization (A, F), confirmed my colorimetric calcium assay (B, G). Elemental analysis
by energy-dispersive x-ray spectroscopy shows increased calcium content (C, H), decreased carbon-to-phosphorus ratio (D, I), and
decreased carbon-to-calcium ratio (E, J) for EXO-MS functionalized scaffolds compared to unfunctionalized nanofibrous scaffolds
alone. Mineralization of functionalized scaffold constructs is comparable to those treated by exogenous administration of OS-EXO at
1 µg/mL and 5 ug/mL over the course of three weeks. * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001.
Fig. 8. Histology of subcutaneous implants. Functionalized scaffolds implanted subcutaneously
for eight weeks shows that cells infiltrate the construct well and occupy the interconnected
spherical macropores (Hematoxylin and Eosin) equally well among nanofibrous scaffolds,
scaffolds loaded with blank EXO-MS, and scaffolds functionalized with OS-EXO-MS. Masson’s
trichrome staining shows increased extracellular matrix deposition in OS-EXO-MS
functionalized scaffolds; von Kossa staining shows increased calcium deposition by calcium
nodule formation in the OS-EXO-MS functionalized scaffolds. Immunohistochemical analysis
for bone sialoprotein (BSP) shows increased BSP protein level as a result of OS-EXO-MS
loaded scaffolds. Scale: 100 µm.

42
Fig. 9. EXO-MS functionalized scaffold leading to bone regeneration in vivo 8 weeks after
implantation without the transplantation of exogenous cells. Hematoxylin & Eosin staining and
Masson’s trichrome staining indicate marrow-containing bone tissue formation as a result of OS-
EXO treatment, in contrast to minimal bone formation in untreated controls (A). Micro-
computed tomography (µCT) of calvaria at 8 weeks (B; scale: 5 µm) indicates significant bone
filling in OS-EXO-MS functionalized scaffold, compared to control scaffold and blank MS
loaded scaffold (C). There is no significant difference in bone mineral density in the mineralized
areas among the groups (D). * p < 0.05, ** p < 0.01

Graphical abstract

43
44