Oral Progenitor Cell Line-Derived Small Extracellular Vesicles as a Treatment for Preferential Wound Healing Outcome

Abstract Scar formation during wound repair can be devastating for affected individuals. Our group previously documented the therapeutic potential of novel progenitor cell populations from the non-scarring buccal mucosa. These Oral Mucosa Lamina Propria-Progenitor Cells (OMLP-PCs) are multipotent, immunosuppressive, and antibacterial. Small extracellular vesicles (sEVs) may play important roles in stem cell–mediated repair in varied settings; hence, we investigated sEVs from this source for wound repair. We created an hTERT immortalized OMLP-PC line (OMLP-PCL) and confirmed retention of morphology, lineage plasticity, surface markers, and functional properties. sEVs isolated from OMLP-PCL were analyzed by nanoparticle tracking analysis, Cryo-EM and flow cytometry. Compared to bone marrow–derived mesenchymal stromal cells (BM-MSC) sEVs, OMLP-PCL sEVs were more potent at driving wound healing functions, including cell proliferation and wound repopulation and downregulated myofibroblast formation. A reduced scarring potential was further demonstrated in a preclinical in vivo model. Manipulation of OMLP-PCL sEVs may provide novel options for non-scarring wound healing in clinical settings.


Introduction
Wound healing and the ensuing scarring place a significant burden on affected individuals as well as on society.In 2014 the global annual cost for wound care was $28.1 billion with a 2018 market research report predicting that the wound-closure market will exceed $15 billion by 2022 and the advanced wound care market exceeding $22 billion annually by 2024. 1 This significant worldwide burden of wound healing, and complications thereof opens up significant opportunities to treat wounds early and, in turn, develop therapies that could reduce the long-term effects of scarring.
In recent years stem cell therapies have shown significant advancements in promoting skin wound healing with results progressing closer to the ultimate aim of delivering scarless wound repair. 2,3Mesenchymal stromal cells (MSCs) have been most commonly studied and have demonstrated significant potential in accelerating the wound healing response and reducing scar formation. 4However, the isolation process of MSCs is often invasive and isolated cells demonstrate significantly reduced potency after only a couple of passages in culture, 5 thereby restricting the clinical development and utilization of such cells.Further to this previous investigations using MSCs have demonstrated an increased risk of tumorigenesis, 6 low engraftment due to their short-lived viability post-injection 7 and difficulties in scalable manufacture. 8ral mucosa lamina propria-progenitor cells (OMLP-PCs) are a novel human adult stem cell population that can be easily isolated from the lamina propria with little to no resultant scar formation. 9,10These clonally expanded cells are CD34− CD45− and CD90+ CD105+ CD166+. 11Like MSCs, this multipotent adult stem cell population is capable of differentiating down mesenchymal lineages (osteogenesis, chondrogenesis, and adipogenesis) as well as down neuronal lineages 11 however, they resist myogenic differentiation. 12OMLP-PCs also possess potent immunosuppressive properties in vitro via both contact-dependent and independent mechanisms. 13Secretions from these OMLP-PCs have also been demonstrated to possess antibacterial activities against Gram-positive and Gram-negative organisms via the secretion of osteoprotegerin and haptoglobin 14 further speaking to the functionality of these cells in the stable maintenance of oral mucosal tissue and also the importance of the OMLP-PC secretome.
][17] Of the paracrine factors secreted by implanted stem cells, small extracellular vesicles (sEVs), commonly referred to as exosomes, have been demonstrated to play an important role in the regenerative potential of their parent cells.sEVs are nanosized particle vesicles (30-130 nm in diameter), believed largely of endosomal origin and are released into the extracellular space when multivesicular bodies fuse with the plasma membrane.sEVs contain a number of biologically active factors such as mRNA, miRNA, proteins and lipids that can target a myriad of biological signaling pathways.sEV based therapeutics demonstrate significant advantages over the previous cell-based therapies.One major advantage is the reduced immunogenic potential; likely due to a lower abundance of immunogenic transmembrane proteins such as MHC complexes. 18The lack of EV replication is also of significant advantage as it presents a reduced risk of tumor generation post-therapeutic intervention.Not only do sEV-based therapeutics replicate the therapeutic benefits of the parental cells 19 but sEVs can also easily be characterized 20 and can even be stored at room temperature following lyophylization 21 resulting in a beneficial therapeutic candidate when compared to a cell-based treatment.An ideal sEV candidate for use as a scarless wound healing treatment should be able to reduce the inflammatory response (Increase in IL-10 secreting M2 macrophages and reduce numbers of M1 macrophages), inhibit the formation of myofibroblasts and modulate collagen deposition. 22n this study, we have immortalized a clonal population of OMLP-PCs (OMLP-PC L ) and demonstrated the functional similarity of this cell line in comparison to the parental cells.From this OMLP-PC L source, we have isolated and characterized sEVs and explored their performance in a range of wound-healing relevant assays.OMLP-PC L sEVs were demonstrated to stimulate positive wound healing responses, when compared to sEV isolated from more traditional MSCs, by increasing fibroblast proliferation and migration rates as well as stimulating potential anti-scarring effects both in vitro and in vivo.The data support further developments of OMLP-PC-derived sEV as therapeutic agents in wound healing settings.

OMLP-PC Culture
Human OMLP-PC cell strains were isolated and cultured as previously described in serum-containing medium (SCM; Dulbecco's modified Eagle's medium (DMEM)) supplemented with 2 mM L-glutamine and antibiotics/ antimycotics (100 U/mL penicillin G, 100 μg/mL streptomycin sulfate and 0.25 μg/mL amphotericin B; Life Technologies, Paisley, UK) and 10% (v/v) fetal bovine serum (FBS-Lot 08G109K & 08Q1072RK) at 37 °C in a 5% CO 2 humidified atmosphere. 11For experimental procedures cells were passaged using 0.05% (w/v) trypsin/0.53mM EDTA.Population doubling levels (PDLs) were calculated as previously published. 23Bone marrow-derived MSCs were purchased from RoosterBio (MD, USA) and cultured in the same way as OMLP-PCs.MSCs were only used up until PDL 20 (Passage 5).For the generation of conditioned medium (CM), FBS was replaced with sEV-depleted FBS (sEV-FBS) that was produced in house by centrifugation at 100 000 × g for 18-20 h at 4 °C in a Beckman Coulter XP100 ultracentrifuge with a 70Ti fixed angle rotor before being passed through both a 0.2 μM and 0.1 μM filter (MerkMillipore; Massachusetts, USA).sEV-FBS was stored at −20 °C until required. 24

OMLP-PC hTERT Immortalization
OMLP-PC cell strains were hTERT immortalized according to the method previously described, 25,26 a technology that we have demonstrated to prolong cellular lifespan by at least 120 PDs past the respective cell strain senescence (relative loss of cell proliferation in non-immortalized cell strains). 27he pBABE amphotropic retrovirus vectors derived from a pLXSN construct and packaged in PA317 cells 28 were kindly provided by Denise Galloway, Seattle, Washington.The pBABE-hTERT-puro encoding retrovirus was constructed by cloning the EcoRI insert of pGRN121 29 into the retroviral vector pBABE-puro. 30Both pBABE-hTERT-puro and pBABEpuro (Mock) vectors were packaged in ψCRIP cells. 25,31Viral supernatants were cultured with cells for 24 h.After 48 h, non-adherent cells were removed, and the remaining cells returned to normal culture conditions under the selection pressure of 0.5 μg/mL puromycin supplemented into the medium.Overall 3 OMLP-PC cell strains were immortalized.From the 3 immortal lines, one line was chosen to be taken forward based on the phenotypic activities that most closely mirrored the activities of the primary cells from which they are derived (Supplementary Fig. S1).All subsequent experiments using this chosen OMLP-PC L were conducted between PDLs 90 and 130.

Flow Cytometry
Cells for flow cytometry were cultured and processed as previously described. 11,13In brief OMLP-PC lines were stained with either CD34-fluorescein isothiocyanin (FITC) (Mouse, clone AC136), CD45-FITC (Mouse, clone 5B1), CD90-FITC (Mouse, clone DG3), CD105-FITC (Mouse, clone 43A4E1) or CD166-FITC (Mouse, clone REA442) (all 10 ug/mL; Myltenyi Biotec), washed twice with PBS before fixing in 4% (v/v) paraformaldehyde (PFA) for 10 minutes on ice.Cells were analyzed on a FACSCanto II flow cytometer equipped with a 488 nm and 633 nm laser excitation source with a minimum of 10 000 events recorded per sample.All data were analyzed using FlowJo Version vX0.7.Data is presented from one biological repeat however, routine cell characterization was performed with consistent results.

RNA Extraction and PCR
RNA was extracted using the Quiagen mini RNA isolation kit.cDNA was generated from 0.5 μg of total RNA using random hexamer primers and MM-LV Reverse transcriptase (Promega UK Ltd, Southampton, UK) as per the manufacturer's instructions.Standard PCRs were carried out using the Platinum Blue PCR supermix (Fisher Scientific; MA, USA) as per the manufacturer's instructions with an initial denaturation step at 94 °C for 5 minutes, followed by 40 cycles at 94 °C (1 minutes), Ta at 55 °C (30 s), 72 °C (1 minutes) and a final elongation step at 72 °C for 10 minutes.cDNA was analyzed for the expression of hTERT [F 5ʹAGAGTGTCTGGAGCAAGTTGC3ʹ and R 5ʹ CGTAGTCCATGTTCACAATCG 3ʹ] Ta 55 °C, and the housekeeping gene glyceraldehyde 3-phosphate dehydrogenase (GAPDH) [F 5ʹCCTCTGACTTCAACAGCGACAC 3ʹ and R 5ʹ TGTCATACCAGGAAATGAGCTTGA 3ʹ] Ta 60 °C.PCR products were separated on agarose gels and visualized using 0.005% (v/v) ethidium bromide.

OMLP-PC L Differentiation and Staining
Osteogenic differentiation was performed utilizing a StemPro osteogenic differentiation kit (ThermoFisher, UK) as per the manufacturer's instructions.OMLP-PC cell lines and cell strains were seeded at a density of 5 × 10 3 cells/well.Cells were cultured under normal conditions for 24 h before the medium was removed and replaced with an osteogenic induction medium.The medium was changed every 48 h for 14 days.After 14 days cells were fixed with 4% (v/v) PFA for 20 minutes before staining with 40 mM alizarin red staining solution for 5 minutes.Cell monolayers were extensively washed with ddH 2 O. Images were obtained using a brightfield inverted microscope.Data is presented from one of 3 biological repeats.
Adipogenic differentiation was performed utilizing a StemPro adipogenic differentiation kit (ThermoFisher, UK) as per the manufacturer's instructions.OMLP-PC cell lines and strains were seeded at a density of 3 × 10 4 cells/well.cultures were re-fed every 48 h for 14 days.After 14 days of culture, cells were washed with PBS and fixed with 4% (w/v) PFA for 20 minutes.Cells were stained with both LipidTox green lipid stain (1:100; Fisher Scientific, UK) and 10 μM Hoechst 33342 in PBS for 30 minutes.Cells were imaged using a Leica SP5 confocal microscope.Data is presented from one of 3 biological repeats.
The colony-forming units (CFUs) of each bacterium were previously calculated from 10 mL (P.mirabilis) or 20 mL (S. pyogenes) overnight cultures derived from single colonies.Cultures were spiral plated using a Whitley automatic spiral plater (Don Whitley Scientific, Shipley, U.K., http://www.dwscientific.co.uk) onto appropriate agar after 23 h at 37 °C (and 5% CO 2 for S. pyogenes).The bacterial colonies were subsequently counted in accordance with the manufacturer's instructions, and the dilution for 100 CFU was determined for each culture.
A total of 100 CFU of either S. pyogenes or P. m i rabilis in PBS were added to 90 μL of OMLP-PC L conditioned medium (CM; 72 h culture in RPMI supplemented with 10% (v/v) FBS, 2 mM L-glutamine and 10% (v/v) Brain, Heart Infusion [BHI]).Cocultures were incubated at 37 °C in 5% CO 2 for 16 h.The bacterial cultures were serially diluted onto TSA (S. pyogenes) or CLED (P.mirabilis) agar plates and incubated at 37 °C overnight before colony counting according to the manufacturer's instructions.An unconditioned medium acted as a control.Data is presented as the mean ± SEM from n = 3 biological repeats.

OMLP-PC L Immunosuppression Assay
Peripheral blood mononuclear cells (PBMCs) were freshly isolated from healthy volunteers who had given informed consent after full ethical approval (Cardiff University Biobank; www.cardiff.ac.uk/biobank).PBMCs were isolated from the buffy coat following density gradient centrifugation on Histopaque.Isolated PBMCs were resuspended and washed 3 times in RPMI medium containing 10% (v/v) FBS to remove any platelets or contaminating histopaque.
As part of the assay, to prevent the OMLP-PC cell lines from proliferating, they were irradiated with 20 Gy (iOMLP-PC L ). 1 × 10 5 freshly isolated PBMCs were seeded into each well of a 96 well plate, stimulated with 500 U/mL of Interleukin 2 (IL-2; R&D Systems, Minneapolis, USA) and either 1000 (1% responder to stimulator cells) or 100 (0.1% responder to stimulator cells) irradiated OMLP-PC L s were added.Following 3 days of culture.A total of 0.5 µCi/well 3 H-thymidine (GE Healthcare, USA) was added to each well for 8 h before the plates were frozen at −20 °C for at least 24 h.After thawing, 3 H-labeled cellular material was harvested onto filtermats (PerkinElmer) and activity was counted using a Wallac 1450 MicroBeta-TriLux 3 Detector (PerkinElmer).Data is presented as the mean ± SD from n = 6 technical repeats.

sEV Isolation
Conditioned medium (CM) was prepared from both OMLP-PC L and MSCs.Cells were seeded into T175 flasks at a density of 4000 cells/cm 2 in 25 mL sEV depleted-SCM.After 48 h, when the cells were approximately 50-60% confluent, the medium was replaced with fresh 25 mL fresh sEV free-SCM and returned to culture for 72 h to condition the medium, (by which point the cells had reached 95-100% confluence).
OMLP-PC L or MSC sEVs were isolated from CM using an ExoSpin kit (Cell Guidance Systems, Cambridge, UK) according to the manufacturer's instructions.In brief, sEVs and proteins were precipitated by mixing with "buffer A" overnight at 4 °C.The mixture was then centrifuged at 16 000 × g for 1 h at 4 °C to pellet vesicles.After 1 h the supernatant was discarded and the pellet was resuspended in 1 mL of PBS.ExoSpin midi columns were equilibrated with 2 × 10 mL of PBS and allowed to drain under gravity.Once all the PBS had passed through the column sEVs were isolated from fractions numbered 7-13.These were determined in pilot experiments where the bulk of sEV-related markers are eluted, as described. 32For the in vivo studies fractions 17-23   were also collected to act as an sEV depleted OMLP-PC L control.Relevant fractions were pooled and concentrated with an Amicon Ultra-15 100 kDa filter to a final volume of between 500 and 1000 μL.A full elution profile of each fraction is demonstrated in Supplementary Fig. S2.
We have submitted all relevant data of our experiments to the EV-TRACK knowledgebase (EV-TRACK ID: EV200030). 33V Nanoparticle Tracking Analysis (NTA) NTA was carried out using a NanoSight LM10 equipped with a blue laser, a CCD camera, and a syringe pump system (Malvern Instruments, UK).Prior to sEV analysis, instrument performance was assessed using 100 nm latex beads (Malvern Instruments).Samples were diluted using nanoparticle free water (Fresenius Kabi, Runcorn, UK) so that the particle concentration was within the linear range of the instrument.Samples were administered and recorded under controlled flow from a syringe pump (set to a speed of 50 arbitrary units).Six replicate videos of 30 s were recorded along with sample temperature for each specimen.Videos were batch analyzed using the integrated NTA 3.1 software, with the cameras sensitivity and detection threshold set to between 14-16 and 1-3, respectively.Particle number was calculated based upon the area under the histogram and modal and mean particle sizes were determined.Background measurements of culture media that had not been exposed to cells and PBS used to dilute sEVs contained negligible particles.

Cryo-electron Microscopy
sEVs were directly adsorbed onto glow-discharged holey carbon grids (QUANTIFOIL, Großlöbichau, Germany).Grids were then blotted at 95% humidity and rapidly plunged into liquid ethane with the aid of a VITROBOT (Maastricht Instruments BV, Maastricht, The Netherlands).Vitrified samples were imaged at liquid nitrogen temperature using a JEM-2200FS/CR Transmission Electron Microscope (JEOL, Tokyo, Japan), equipped with a field emission gun, and operated at an acceleration voltage of 200 kV.

Flow Cytometry of sEV-Coated Microbeads
A total of 25 μg of sEVs were bound to 20 μL of CD63 coated Dynabeads (ThermoFisher, USA) in a 100 μL final volume overnight at 4 °C made up of 0.1% (w/v) sterile filtered Bovine Serum Albumin (BSA).Following an overnight incubation, beads and sEVs were captured using a magnet and washed with 0.1% BSA solution.After resuspending beads and sEVS in 100 μL 0.1% (w/v) BSA, either the IgG control or test antibodies were added; IgG1 FITC (50 μg/mL), IgG1 PE (50 μg/mL), CD81 FITC (50 μg/mL), CD63 FITC (50 μg/ mL) and CD9 PE (50 μg/mL), (all from Miltenyi, Germany).Primary antibodies were incubated at 4 o C in the dark for 30 minutes.Samples were analyzed on a FACSCanto II Flow cytometer (BD Biosciences) equipped with a 488 and 535 nm laser excitation source a minimum of 10 000 events were recorded per sample.Fluorophore-conjugated immunoglobulins were used as controls.All data were analyzed using the software package FlowJo version vX0.7.
sEV Uptake Experiments SEV uptake into recipient fibroblasts was carried out as previously described. 34Cells were seeded into either 6 well plates at 50 000cells/well (for flow cytometry) or 8-well chamber slides at 4500 cells/well (fluorescent microscopy) and left for 48 h to adhere and proliferate in SCM.After 48 h SCM was removed and cells switched to serum free medium (SFM) for 48 h.Following this 48 h incubation, 200 μg/mL of C 5 -maleimide-Alexa633 was added to 150 μL of sEVs and incubated at room temperature in the dark for 1 h.Labeled exosomes were then added to prepared exosome spin columns (Invitrogen) and centrifuged at 750xg for 3 minutes to allow exosomes to pass through the resin and any unbound dye to remain in the resin.
Labeled exosomes were diluted in SFM to a concentration of either 10 μg/mL or 50 μg/mL and incubated with cells for 1 h.After 1 h, exosome labeled cells were washed twice with PBS before proceeding to either flow cytometric analysis or confocal laser scanning microscopy.Either 10 000 events were recorded per sample for flow cytometry or cells were counterstained with 0.8 nM Phalloydin-Atto-594 and DAPI before imaging on a Zeiss LSM880 Airyscan microscope.

In Vitro Proliferation Assay
Dermal fibroblasts were seeded at 5 × 10 3 cells/well into 96 well plates in 100 μL SCM and allowed to adhere for 24 h.SCM was then removed and replaced with 100 μL SFM for 24 h to growth arrest the cells.Following growth arrest, 100 μL of medium was added to cells containing sEVs at either 0, 10, 50, or 100 μg/mL sEVs.Cells were returned to culture for either 24 h (day 1) or 72 h (day 3) after which Orangu-solution (Cell Guidance Systems) was added (10 μL) into each well and cells returned to the incubator for 4 h.After 4 h the absorbance was measured at 450 nm using a microplate spectrometer (FLUOstar Omega; BMG Labtech; Aylesbury, UK).Data are presented as the mean ± SEM from n = 3 biological repeats.

In Vitro Cell Migration Assay
A total of 1.5 × 10 3 dermal fibroblasts were seeded into each side of a 2-well silicone Ibidi cell culture insert (Thistle Scientific; Glasgow, UK) inserted into a well of a 24-well plate and incubated at 37 °C 5% CO 2 for 48 h until the cells formed a confluent monolayer.After 48 h, the silicone insert was removed and 400 μL of SFM containing sEVs at either 0, 10, 50, or 100 μg/mL was added.Plates were placed into a Cell-IQ time lapse microscope (CM Technologies; Elmshorn, Germany) for 48 h during which time images of each well were acquired every 30 minutes.Migration was analyzed using the integrated analysis software within the Cell-IQ microscope.Individual cells were tracked (10 cells per well) and their cell migration speed was calculated (distance migrated/ time).Data are presented as the mean ± SEM from n = 3 biological repeats.

In Vitro Inhibition of Myofibroblast Formation
Dermal fibroblasts were seeded at a density of 5 × 10 3 cells/ cm 2 in SCM.After 24 h SCM was removed and replaced with SFM for a further 24 h to growth arrest cells.After 24 h, SFM was removed and replaced with SFM containing 1 ng/mL Transforming growth factor (TGF) β1 only, 1 ng/mL TGFβ1 and sEVs (concentrations 100, 50, or 10 μg/mL), or SFM only for 72 h.After 72 h, cells were either used for protein extraction or fixed and stained for immunohistochemistry as previously described. 11,13Fluorescence images were acquired using 1 μg of primary αSMA monoclonal antibody (Clone 1A4, Abcam; Cambridge, UK).A total of 1 μg of AlexFlor488 secondary antibody (Abcam) was then added for 1 h, washed and fluorescent images was obtained using a Zeiss Airiscan880 laser scanning confocal microscope.Data are presented as the mean ± SEM from n = 3 biological repeats.

Western Blot
Dermal fibroblasts were cultured as described above in 6 well plates.Following myofibroblast differentiation cell monolayers were washed with PBS before harvesting in lysis buffer (RIPA Buffer Kit; Santa Cruz Biotechnology) containing a protease cocktail inhibitor mix.A total of 10 μg of cell lysate was separated by SDS-PAGE and transferred to a PVDF membranes.Membranes were blocked overnight in 5% (w/v) non-fat powdered milk in PBS containing 0.5% (v/v) Tween 20 (PBST).Membranes were probed with primary antibodies; mouse anti-human alpha smooth muscle actin (1 μg/mL; Clone 1A4, Abcam, UK) for 2 h at RT. Membranes were washed 6 times in PBST before incubation with a secondary goat anti-mouse HRP conjugated antibody for 2 h at RT. Membranes were washed 6 times in PBST before being analyzed with Amersham -ECL Plus.HyperFilm X-ray film (GE Healthcare) was then placed onto the membrane and left to expose for a predetermined time.The film was developed using a Curix 60 X-ray developer (AGFAGevaert, Belgium).Each experiment was carried out alongside 10 μl of Sea Bule-Plus 2 Precision stain and 3 μL of MagicMark-XP ladder (Both LifeTechnologies, USA).To normalize protein loading, membranes were probed for beta Actin.Relative intensity was calculated and normalized to a no sEV control.

Animals
A total of 8-12 week old male C57/B6/J mice were purchased from Charles river UK (Margate, UK).Mice were housed at the JBIOS unit at Cardiff University and maintained at 20 °C on a 12-h light/dark cycle, with free access to standard chow and water.All procedures were reviewed by a local ethics committee and were performed under home office project and personal licences (PPL: 302362).

In Vivo Wound Model
Mice were anesthetized by 3-3.5% isoflurane delivered in 2 L per minutes in 100% oxygen.Once areflexic, mice received a sub-cutaneous dorsal injection with 10 μL Temgesic/ Buprenorphine (equivalent to 1 μg Temgesic).The mice were then shaved and re-tested to ensure the areflexic state was maintained.Ear clips were conducted to identify mice then a spinal midline was drawn on each mouse before the mouse was rotated to its side, the skin folded using the spinal midline as a reference point and 2 4 mm punch biopsy needles (Farla Medical Ltd, UK) were used to create 4 wounds (2 wounds per biopsy needle).Any skin remaining in the wounds was trimmed and removed using sterile scissors.The wounds were not dressed.Mice were then transferred to a recovery warming box until consciousness was regained before being transferred into its cage lined with a paper towel.Mice were monitored at 1 and 3 h post wounding and again before the end of the day.24 h post wound the mice were then transferred into their original home cage (with woodchips/sawdust nesting materials present).
Immediately post wounding (day 0) each mouse received 50 μL of treatment, which consisted of either no treatment or topically administered 100 μg/ml of OMLP-PCL sEVS or 100 μg/mL of an sEV depleted OMLP-PC L treatment.
With 4 wounds created per mouse, each mouse received one of each treatment per wound, n=4 mice.On day 3 post wounding mice were again anesthetized briefly (less than 5 min) and 50 μL of either 2 treatments or PBS were again applied to the same wounds.Mice were then allowed to recover.
Mice were humanely killed (day 7 post wounding) via CO 2 asphyxiation as set out in the humane killing of animals under schedule 1 of the animals (scientific procedures) act 1986.The wound sites and a small area of surrounding tissue for each mouse were dissected.Tissue samples were placed in 4% paraformaldehyde for 24 h, then 70% methanol for 24 h then the tissue was processed for paraffin wax embedding and sectioned (10 μm).

Collagen Staining
Staining of collagenous fibers in sections was undertaken using the Masson trichrome stain (Abcam ab150686) and visualized by bright-field microscopy.

Image Acquisition and Analysis
Images were captured with wither a Leica microscope (Leica DM 2000 LED, with a Leica DMC 2900 camera) or epifluorescent (Evos, Thermo Fisher scientific) microscope.Brightness/contrast was adjusted to the same extent for all image acquisition and image analysis was always undertaken on non-processed (RAW) images.In terms of quantitation, high-power images were used, with 3 fields of view (wound center, left, and right) used.Images were processed and analyzed in ImageJ (Version 1.49u).For collagen stained images, color threshold was applied to the images using scale color 70-130, to remove all colors from the image save the blue collagen color, the mean gray intensity of blue pixels was determined.For fluorescent image, analysis mean gray intensity of Texas red (Alexafluor 568) channel images were obtained.Data were then analyzed with a One Way Anova with a Bonferroni post hoc test with P < .05considered statistically significant.

Ethical Approval
Ethical approval was provided by West of Scotland REC 5, United Kingdom.

Statistical Analysis
All statistical analyses were performed using GraphPad Prism version 7.00 for Mac (GraphPad Software, La Jolla California USA, www.graphpad.com).Data was assessed for normality using a Shapiro-Wilk test for normality and if Gaussian distribution was confirmed mean values was performed using one-way analysis of variance (ANOVA) with post hoc Tukey test (equal variances assumed) or a Games-Howell test (unequal variances determined).Variance analysis was performed using Levene's test.Statistical significance was assumed at P < .05.No data were excluded from the analysis.

Results
hTERT Immortalized OMLP-PCs Retain Their Progenitor Cell Properties/Functionality OMLP-PC cell lines (OMLP-PC L ) cultured consistently in a selection medium containing 0.5 μg/mL puromycin, demonstrated a typical fibroblast-like morphology (Fig. 1A) thereby confirming the positive infection of the pBABE-hTERT construct into the recipient cells.This was also confirmed by RT-PCR which demonstrated expression of the hTERT gene (Fig. 1B).Long term culture of the cell line and a patientmatched mock-infected cell strain (ie, cells transfected with the empty vector-no HTERT added) demonstrated the cell line was capable of continued proliferation while the strain underwent cell senescence and stopped dividing as could be observed by a plateauing of the growth curve (Fig. 1C).Flow cytometry demonstrated that the oral cell line was positive for the progenitor cell markers CD90, CD105, and CD166 whilst negative for fibrocyte and hematopoietic markers CD34 and CD45 (Fig. 1D).Overall this suggested that the progenitor cell strain had been successfully immortalized.
Functionally, this immortalized OMLP-PC L demonstrated plasticity, undergoing adipogenesis and osteogenesis (Fig. 2A) as previously published for the corresponding cell strain. 11Additionally, the OMLP-PC L retained its secretory antibacterial properties, inhibiting bacterial growth of both Gram-positive and Gram-negative bacteria (Fig. 2B; P < .01).Furthermore, the OMLP-PC L demonstrated an ability to significantly (P < .0001)inhibit T-cell proliferation when PBMCs were stimulated with IL-2 (Fig. 2C).It was demonstrated that this inhibitory effect was not dosedependent, something that was previously also demonstrated for the OMLP-PC cell strains. 13Collectively all this information suggests that OMLP-PCs can be immortalized and still retain their original progenitor cell properties and functionality.
sEVs Can Be Isolated From Cell Conditioned Medium OMLP-PC L and MSC CM were used to generate sEVs.NTA analysis demonstrated sEVs can be purified/enriched from both OMLP-PC L and MSC cultures (Fig. 3A, 3B) and are well within the expected size range for sEVs of 30-130 nm.Images obtained from Cryo-EM confirmed a heterogeneous population of vesicles exists, with a rounded structure and a distinctive, outer lipid bilayer indicative of sEVs for both OMLP-PC L and MSCs (Fig. 3C, 3D).Other membranous structures were apparent in the preparations as rare events and similar to other studies of sEVs. 35Tetraspannin expression was determined using flow cytometry of sEVs following capture on CD63 coated dynabeads.Using a CD63 bead capture eliminates any misleading signal that could have arisen due to soluble tetraspannins or membrane fragments.Both OMLP-PC L and MSC sEVs demonstrated positive expression of CD9, CD81, and CD63 following a CD63 bead capture confirming the co-localization of these markers within the specimen, likely in the form of sEVs (Fig. 3E, 3F).Analysis of cell markers CD90 and CD105 were negative on purified sEVs following the same CD63 bead capture (Supplementary Fig. S3).After floatation on a continuous sucrose gradient purified sEVs from OMLP-PC L and from MSCs were found to have a buoyant density between 1.140-1.260g/cm 3 and 1.175-1.197g/cm 3 respectively based upon tetraspannin expression (Supplementary Fig. S4).Both OMLP-PC L and MSC-derived sEVs labeled with a 488 Alexa Fluor Malemide dye demonstrated uptake into dermal fibroblasts, and into perinuclear puncta consistent with endosomes after a 1-h incubation (Fig. 3G, 3H).Individual image channels can be seen in Supplementary Fig. S5 as well as a dye only control that demonstrates fluoresce in the 488 (sEV labeled) channel must be due to binding to sEVs and not because of free unbound dye being internalized by cells.This was also confirmed by   flow cytometry demonstrating a heterogeneous uptake across a large population of cells (Supplementary Fig. S6).

OMLP-PC L sEVs Demonstrate Positive Wound Healing Capabilities
OMLP-PC L and MSC sEVs were next examined in respect of their wound healing functional capabilities in vitro.Firstly, we explored the effects on fibroblast proliferation using a colorimetric assay.Both sEV types demonstrated an ability to increase fibroblast proliferation by day 3 in a dose-dependent manner.Incubation with either 100 μg/mL or 50 μg/mL sEVs, significantly increased fibroblast proliferation (Fig. 4A; P < .001).However, at the lower concentration of 10 μg/mL only, OMLP-PC L sEVs retained their function to increase proliferation (Fig. 4A; P < .001)suggesting a greater potency when compared to MSC-derived sEVs.To further support the greater potency of OMLP-PC L sEVs, significant differences were also observed at the higher, 100 μg/mL concentration between OMLP-PC L and MSC sEVs (Fig. 4A; P < .05).
We next explored potential effects on fibroblast motility, using an automated time-lapse microscopic method.Average cell migration speeds were calculated and it was clear that at 100 μg/mL both OMLP-PC L and MSC sEVs stimulated fibroblasts to migrate at a significantly faster speed (Fig. 4B; P < .001),compared to controls but with OMLP-PCL sEVs stimulating fibroblast migration speeds significantly above treatment with MSC sEVs (Fig. 4B; P < .0001).However, at the lower concentration of 50 μg/mL only, OMLP-PC L sEVs demonstrated an ability to significantly increase fibroblast migration speeds (Fig. 4B), significantly above the migration speeds of MSC treated cells (P < .001)again suggesting a superior potency of these sEVs in this assay.

OMLP-PC L sEVs Demonstrate Potential Antiscarring Capabilities
One of the major scar-forming cells is the myofibroblast, which is routinely identified through their positive expression of αSMA.When dermal fibroblasts are cultured with TGFβ1 they polymerize the α-smooth muscle-specific isoform of actin into stress fibers.The onset of such αSMA-stress fibers is a defining hallmark of myofibroblast formation.
When dermal fibroblasts were cultured with TGFβ1 together with sEVs from OMLP-PC L and MSCs, the amount of αSMA positive cells was significantly reduced, as determined by both ICC and Western blotting (Fig. 4C, 4D; P < .001).Interestingly, when OMLP-PCL sEVs at 100 μg/mL was used in the assay, the amount of αSMA produced was not significantly different from the amount of αSMA produced by fibroblasts that had not been cultured with TGFβ1-ie, levels were reduced to baseline levels (P > .05;Figs.4C, 4D).This was not the case for MSC sEVs with OMLP-PCL sEVs demonstrating a significantly higher potency at both 100 μg/ mL and 50 μg/mL (P < .001).Furthermore, at 50 μg/mL, only OMLP-PCL sEVs was capable of significantly reducing the amount of αSMA present in the TGFβ1 treated fibroblasts (P < .01;Fig. 4D).

In Vivo Analysis Demonstrates Anti-scarring Characteristics of OMLP-PC L sEVs
Based on the in vitro findings, OMLP-PC L sEVs were also analyzed for their ability to influence scar formation in an in vivo wound healing model to demonstrate real-world proof of concept.Whilst there was no effect on the wound size on day 4 (Fig. 5A, 5B), mice treated with the OMLP-PC L sEV rich treatment (Fig. 5C, 5E) showed a significant (P < .05)decrease in dermal production of αSMA, suggesting a reduction in myofibroblast number, whilst mice treated with an sEV depleted protein treatment showed no significant reduction (P > .05) in the dermal staining of smooth muscle actin compared to non-treated controls (Fig. 5C, 5E).Masson's trichrome staining revelated that mice treated with the OMLP-PC L sEV rich treatment (Fig. 5D) showed significantly (P < .01)reduced dermal collagen deposition, but mice treated with an OMLP-PC L sEV depleted fraction showed no significant alteration in collagen deposition when compared to non-treated controls (Figs.5D, 5F).All wounds and lowpower mason trichrome staining images are presented in Supplementary Figs.S7-S9.

Discussion
OMLP-PCs are a relatively newly described adult progenitor cell population that can be easily isolated from the buccal mucosa. 11,13,14Although these progenitor cells demonstrate a significant increase in their in vitro lifespan compared to bone marrow mesenchymal stromal cells (which generally senesce in a little over 30 PDs 36 ) they still undergo telomere dependent senescence (after approximately 60 PDs).The scale up of these clonal oral cells to a clinically relevant level for the isolation of soluble factors such as exosomes/ sEVs would thus result in cells approaching or reaching their point of senescence.Hence, in terms of their exploitation as a source for soluble factors for potential regenerative medicine applications, they are limited.However, if there was the capability to immortalize such cells and in so doing not affect their progenitor cell phenotype then this may allow for prolonged periods of cell culture and a sustainable and consistent supply of therapeutic factors. 37n this study, we have demonstrated the hTERT immortalization of an easily accessible clonal progenitor cell population and the effects of immortalization on progenitor cell phenotype and function in vitro.Immortalisation of a starting clonal cell population was important to reduce any heterogeneity between individual cells.Continued expansion in monolayer culture demonstrated that OMLP-PC L was able to grow well past the point where the patient-matched cell strains underwent senescence, thus confirming the immortalization of the OMLP-PC strain.As previously published, OMLP-PCs are characterized by the positive expression of stem markers CD90, CD105, and CD166 and the negative expression of hematopoietic or fibrocyte markers CD34 and CD45 (11).Importantly, not only did the OMLP-PC L mirror these expected markers but they also shared this CD profile with the more extensively studied but less easily accessible and less potent BM-MSCs. 38,391][42] When looking at other published reports, hTERT immortalized MSCs have also been reported to retain their ability to undergo differentiation at similar levels to the parent strains. 38,43Interestingly, both Jun et al 44 and Wolbank et al 38 have also reported that some hTERT immortalized MSC cell lines show an improved adipogenic differentiation capability following immortalization, likely attributed to a decrease in committed cells that spontaneously undergo differentiation resulting in a higher number of cells  capable of undergoing directed differentiation, 45 which is in agreement with our findings.Functionally, hTERT immortalization produced OMLP-PC L that retained their antibacterial and immunosuppressive capabilities.As previously reported the OMLP-PC antibacterial effect is due to the secretion of both haptoglobin and osteoprotegerin therefore it suggests that hTERT immortalization has not affected the production of these antibacterial molecules, however, further experimentation is required to determine if such activity is actually associated with sEVs. 14In this study data from only one OMLP-PC L has been presented, the cell line that fully demonstrated the phenotypic activities of the original primary cells from which they were derived (much like a number of commercial entities working in this space utilize a single "master" cell line to produce all their commercial products).
Although there are many methods available for sEV isolation from cell-conditioned media, we undertook a commercial precipitation-based approach, for the first-ever study of this source of EVs.Precipitation methods are often criticized for their lack of specificity and the co-precipitation of a host of non-vesicular material.However, in this system, the second step relies on size exclusion chromatography where the column is highly efficient at eliminating as much as 97% of the specimen protein that is not related to vesicles. 32sEV isolation techniques have been used by numerous groups who have then reported on the functional effects of those sEVs.For example, sEVs isolated by precipitation based kits have been shown to retain complex functions and can induce osteogenesis, 46 drive angiogenesis, 47 inhibit neointimal hyperplasia 48 and promote wound healing 49 demonstrating that these tractable and simple precipitation-based isolation methods are capable of isolating sEVs which retain their functional properties are of utility in such applications.In future iterations, however, the authors would be keen to eliminate precipitation, and consider other methods of selective concentration such as tangential flow filtration to minimize vesicle complexation or damage and minimize co-isolating material.
Here, the wound healing capabilities of sEVs isolated from OMLP-PC L were compared to sEVs isolated from MSCs.sEVs were assessed in terms of their abilities to positively influence wound healing as well as any potential anti-scarring capabilities.
We report herein, that OMLP-PC L and MSC-derived sEVs were both capable of significantly increasing fibroblast proliferation and cellular migration.In agreement with these findings, stem cell exosomes/sEVs have previously been demonstrated to increase the proliferation of primary fibroblasts 17,49,50 as well as cancer cells. 51Additionally, stem cell derived exosomes/sEVs have been shown to significantly increase cell migration. 49,52However, this is the first report of OMLP-PC sEVs and their demonstrated greater potency/function when compared to MSC-derived sEVs.OMLP-PC L and MSC derived sEVs were both capable of significantly inhibiting the formation of myofibroblasts with OMLP-PC L sEVs again having a more potent effect.Importantly, this modulation of wound healing functionality was directly dependent on the dose of sEVs added into the assay, something which the International Society of Extracellular Vesicles requires to be demonstrated for a potential therapeutic sEV. 20,53Previous studies that have used MSC-derived sEVs have demonstrated a significant reduction in myofibroblast formation through inhibition of the TGFβ/SMAD2 pathway, 54 a possible pathway that could also be inhibited by OMLP-PC L sEVs.Other studies using OMLP-PC-derived sEVs have only focused on their effects on cell proliferation 50 and their impact on myofibroblast formation was previously unknown.It was known, however, that oral fibroblasts, that most probably contain some oral progenitors, are resistant to differentiation into myofibroblasts, 55,56 an effect that could, inherently, be due to OMLP-PC derived sEVs.Importantly, in terms of realworld functionality, in a murine wound model OMLP-PC L sEVs also demonstrated an ability to inhibit the formation of αSMA positive myofibroblasts and significantly reduced collagen deposition 7 days post wounding suggesting that such oral sEVs may potentiate longer-term scarring outcome.8][59] Whilst the method of action was not examined in this study, we hypothesize that the biological function of sEVs could be either due to the transfer of protein, miRNA or a combination of the 2 as has been observed with other stem/progenitor cellderived sEVs. 60,61Future assessment of sEV cargo (protein/ miRNA) would allow for the identification of functional sEV components but also allow for an understanding of the differences observed between the OMLP-PC L and MSC sEVs used in this study.
The wound healing market is diverse with a number of approaches targeted at preventing/modulating scar tissue.These include dressings (eg, silicone sheets), topical/injected agents (eg, herbal extracts, drugs [such as imiquimod, metformin, propanolol], corticosteroids, botox, growth factors, hormones [insulin, erythropoietin]), anti-coagulants, anti-oxidants, pressure therapy, nutrition, laser treatment, radiation treatment and cryotherapy (for recent reviews see 62,63 ).The OMLP-PC L -derived sEVs described within this manuscript could have future potential within the burns market.Incorporation into a delivery material such as a hydrogel would allow for sustained release of the beneficial sEVs directly into the wound site.Whilst the experiments conducted here only test the functionality of OMLP-PC L sEVs in relatively small skin wounds, wounds that heal within 7 days with no treatment in vivo, these proof-of-concept experiments are a prerequisite prior to moving forward to larger scale (porcine) wound models (eg, a burns model).
Overall, due to the increase in potency, clonal nature, and the formation of a consistent/functional cell line, we suggest OMLP-PC L sEVs as a preferential sEV population when compared to the more commonly studied bone marrowderived sEVs.These OMLP-PC L sEVs are a novel and translatable therapeutic product that has the potential to improve the lives of the ever-increasing numbers of people suffering from severe fibrosis and scarring.

Figure 1 .
Figure 1.Following hTERT immortalization, OMLP-PC L demonstrated a normal fibroblast morphology whilst cultured in a puromycin selection medium (A), continued to proliferate past the senescent point of a matched Mock cell strain (cells treated and transfected in an identical way to the immortalized cells except an empty vector was used instead (ie, no hTERT was added to the cells) (B; arrow indicates point of immortalization) and demonstrated successful incorporation of the hTERT gene which was confirmed by RT-PCR (C).Flow cytometry demonstrated the OMLP-PC L retained its progenitor cell surface characteristics being CD90, CD105, and CD166 positive but negative for CD34 and CD45 (D).Scale bar = 100 μm.

Figure 3 .
Figure 3. sEV characterization by both NTA (data from 6 technical repeats) and Cryo-EM demonstrated particles of the expected size for sEVs for both OMLP-PC L (A&C) and MSCs (B&D; Scale bar = 100 nm).Tetraspannin expression was determined by flow cytometry and demonstrated OMLP-PC L sEVs (E) and MSC sEVS (F) were positive for CD9, CD81, and CD63.Fluorescently labeled sEVs (Green-488) from OMLP-PC L (G) and MSCs (H) demonstrated uptake into dermal fibroblasts which were counterstained to visualize actin (Magenta) and the cell nucleus (Blue)-images are confocal slices.Scale bar = 50 μm.Error bars for NTA = SD.

Figure 5 .
Figure 5. Four identical wounds were created on the back of each animal and treatments were added for 4 days (A; I-no treatment, II-PBS, III-sEV depleted treatment, IV-sEV containing treatment).OMLP-PC L sEV treatment on mouse wounds demonstrates no statistically significant effect on wound size at 4 days (B).However, OMLP-PC L sEV treatment results in a reduction in α-SMA immunofluorescence and collagen deposition in vivo.Merged images showing nuclear DAPI (blue) staining and α-SMA (red) staining for non-treated, OMLP-PC L sEV depleted treatment, OMLP-PC L sEV treatment (C) and quantification of fluorescent α-SMA (E).Scale bar = 200 μm n = 4 wounds per group.Masson's trichrome staining reveals collagen deposition in wounds from non-treated, OMLP-PC L sEV depleted treatment, OMLP-PC L sEV treatment (D) and quantification of collagen staining (F).n = 4 wounds per group, *P < .05,**P < .01,Scale bar = 30 μm.