Engineered Vasculogenic Extracellular Vesicles Drive Nonviral Direct Conversions of Human Dermal Fibroblasts into Induced Endothelial Cells and Improve Wound Closure

Abstract

Vasculogenic cell therapies have emerged as a powerful tool to increase vascularization and promote tissue repair/regeneration. Current approaches to cell therapies, however, rely mostly on progenitor cells, which pose significant risks (e.g., uncontrolled differentiation, tumorigenesis, and genetic/epigenetic abnormalities). Moreover, reprogramming methodologies used to generate induced endothelial cells (iECs) from induced pluripotent stem cells rely heavily on viral vectors, which pose additional translational limitations. This work describes the development of engineered human extracellular vesicles (EVs) capable of driving reprogramming-based vasculogenic therapies without the need for progenitor cells and/or viral vectors. The EVs were derived from primary human dermal fibroblasts (HDFs), and were engineered to pack transcription factor genes/transcripts of ETV2, FLI1, and FOXC2 (EFF). Our results indicate that in addition of EFF, the engineered EVs were also loaded with transcripts of angiogenic factors (e.g., VEGF-A, VEGF-KDR, FGF2). In vitro and in vivo studies indicate that such EVs effectively transfected HDFs and drove direct conversions towards iECs within 7–14 days. Finally, wound healing studies in mice indicate that engineered EVs lead to improved wound closure and vascularity. Altogether, our results show the potential of engineered human vasculogenic EVs to drive direct reprogramming processes of somatic cells towards iECs, and facilitate tissue repair/regeneration.

Graphical Abstract

Engineered human EVs loaded with ETV2, FLI1, and FOXC2 (EFF), can effectively drive vasculogenic reprogramming in human dermal fibroblasts and aid in in vivo tissue repair and regeneration. Such EVs have the potential to serve as a powerful nanocarrier platform for vasculogenic factors to treat a myriad of ischemic disorders.

INTRODUCTION

Extracellular vesicles (EVs) are spherical lipid-bilayer nanoparticles secreted by all healthy and diseased cells in the body. EVs originate from the endosome or plasma membrane and are classified into three sub-populations: exosomes, microvesicles, and apoptotic bodies, according to their release mechanism and size^1–3. EVs are capable of packing and transporting payloads of miRNAs, mRNAs, DNA, and proteins among other molecules⁴. The molecular cargo is protected from extracellular degradation by their lipid bilayer membrane. EVs have shown to be implicated in physiological processes like paracrine, autocrine, and immune system signaling⁵. They also serve an important role in surface molecular trafficking and horizontal transfer of DNA, mRNA, and proteins among different cell types⁶. Cells can incorporate the molecular cargo derived from EVs through several mechanisms, including membrane receptor-ligand interactions¹, or direct fusion of the EVs with the membrane of recipient cells via endocytosis. The increase in knowledge on EV composition, biogenesis, and their role in intercellular communication, has given rise to a research field that aims at understanding how EVs could be implicated in diseases onset and propagation^7–10. Additionally, our lab and others have shown their significant potential to serve as efficient nanocarriers for an extensive array of therapeutic applications, including regenerative medicine and cancer^{7, 11–20}.

Ischemic disorders and tissue injury often result in deficient vascularization, which can lead to infections, inflammation, and impaired tissue regeneration (e.g., non-healing ulcers, partial necrosis)^{21, 22}. Cell therapies, involving the use of vascular progenitor cells or induced endothelial cells (iECs), have emerged as a powerful approach to increase neovascularization and promote tissue repair and regeneration under multiple conditions, including myocardial ischemia, limb ischemia, and stroke, among others^{19, 23–26}. iECs provide growth factors, such as vascular endothelial growth factor (VEGF), and iEC-driven neovasculature serves as a scaffolding structure to support and guide cell migration and proliferation, promoting tissue repair^{27, 28}. Current approaches to vasculogenic cell therapies, however, rely on transplantation of progenitor/stem-like cells^{29, 30}, which pose significant risks in terms of uncontrolled/undesired differentiation, tumorigenesis, genetic/epigenetic abnormalities, immunogenicity, and ethical considerations^{31, 21, 28, 32, 33}. Moreover, reprogramming methodologies used to generate induced pluripotent stem cells (iPSCs)-derived endothelial cells (ECs) are heavily dependent on viral vectors, which pose additional biosafety and translational concerns.

Recent studies by us and others have shown that transcription factor genes/transcripts Ets variant 2 (ETV2), Friend leukemia virus integration 1 (FLI1), and Forkhead box C2 (FOXC2) (EFF) can be used to directly reprogram somatic cells into iECs^{18, 19, 28, 34–37}. ETV2 (Ets variant 2) has recently been reported as a pioneering transcription factor, implicated in the embryonic development of endothelial and hematopoietic lineages^{38 39, 40}. ETV2 plays an important role inducing relaxation of the chromatin by the recruitment of factors such as Brahma-related gene-1 (BRG1) and ETS Transcription Factor (ELK3), this recruitment opens the chromatin and increases H3K27ac deposition, resulting in the activation of gene expression³⁸. Recent reports suggest that ETV2 can induce Rhoj expression, and that the ETV2/Rhoj network plays a key role in regulating the migration of early endothelial progenitor cells during embryonic development⁴¹. An essential function of ETV2 is its interaction with forkhead transcription factors, like a FOXC2, via the FOX:ETS motif, which is a composite cis-element. This ETV2/FOXC2 complex can regulate at least 30% of endothelial gene expression⁴². A complete knockdown of ETV2 affects the expression of genes like FLI1, and can lead to embryonic lethality⁴³. FLI1 is a key regulator of hematopoietic progenitor cell differentiation and maintenance, and vascularization during embryonic development⁴³. FLI1 and Ets-related gene (ERG), are implicated in hematopoietic and endothelium development, while FOXC2 is an important regulator of essential angiogenesis processes like cell migration, remodeling, and maturation^{42, 44–46}. On the other hand, growth factors such as FGF2 and VEGF-A synergistically promote angiogenesis, and proliferation of endothelial cells ^{40,47, 48}. Additionally, recent studies suggest that FGF2 is involved in maintenance of vascular integrity in adult vasculature by regulating VE-cadherin/p120-catenin coupling⁴⁹. These VE-cadherins/p120-catenin interactions are essential for maintaining endothelial barrier function and the prevention of endocytosis⁵⁰.

Recently we reported that in vitro cell nanotransfection or in vivo tissue nanotransfection (TNT) with EFF can drive the development of iECs and improve tissue vascularization and repair/regeneration under different models of tissue ischemia^{18, 19, 25, 26}. While in both cases we document the release of EFF-loaded EVs from EFF-nanotransfected cells and tissues, currently no study has looked into the production and characterization of human EFF-loaded EVs for therapeutic applications. Here we show that primary human dermal fibroblasts (HDFs) can be driven to release EFF-loaded EVs via non-viral transfection of expression plasmids for EFF, and that such EVs can induce direct conversion of somatic cells into iECs in vitro and in vivo. Finally, we showed that EFF-loaded human EVs can lead to improved wound closure in mice. Our results highlight the potential use of engineered EVs as a novel cell-free therapy to help overcome many of the limitations of current progenitor/stem cell-based therapies.

RESULTS

EFF transfected human dermal fibroblasts (HDFs) release vasculogenic EVs with tailored cargo

EFF loaded EVs were derived from HDFs after electroporation with plasmids encoding for ETV2, FLI1, and FOXC2 (EFF)¹⁸ (Figure 1A). HDFs transfected with a sham plasmid (i.e., empty vector with the same backbone) were used to derive control EVs. HDF transfection efficiencies ranged between 30 and 40% for the sham (control) and EFF groups, respectively (Supplemental Figure 1). Dynamics of engineered EV release revealed a peak 48 hours after transfection of the donor HDFs, with particle concentrations ranging around 7×10⁹ EVs/mL for EFF-loaded EVs, and 5×10⁹ EVs/mL for control EVs, respectively (Figure 1B). EV size remained consistent over time, around 200 nm for both EFF-loaded and control EVs, as confirmed via nanoparticle tracking analysis by NanoSight, and cryo-transmission electron microscopy (cryo-TEM) (Figure 1C–D and Supplemental Figure 2). Western blot characterization confirmed the presence of the EV marker, CD63, in both EFF and control EV preparations (Figure 1E). Quantitative and absolute reverse transcription polymerase chain reaction (qRT-PCR and aRT-PCR) analysis confirmed robust and sustained upregulation of EFF expression in donor HDFs for 12–48 hours after transfection, with a significant increase at 48 hours for ETV2 and FLI1, and slightly higher expression for FOXC2 at 24 hours (Figure 1F). qRT-PCR characterization of the EVs, on the other hand, shows a significant and sustained increase in EFF loading between 12–48 hours post-transfection, with 48 hours consistently showing one of the highest levels of each factor, thus suggesting that 48 hours post-HDF transfection may represent the best timepoint for EV collection for downstream use (Figure 1G). To determine whether the EVs were loaded with plasmid DNA, intact EVs were treated with DNase to remove potential coprecipitates (Supplemental Figure 3A). Subsequently, the presence of plasmid DNA within the EVs was confirmed by amplification of the poly-linker region of each plasmid via conventional PCR (Supplemental Figure 3B). Moreover, absolute PCR characterization showed significant packing of both plasmid DNA and mRNA content in EFF EVs compared to control EVs (Supplemental Figure 3C–3E). Further characterization of the transcript content packed in the EFF-loaded EVs suggested the presence of additional regulators of vasculogenesis and angiogenesis, specifically VEGF-A and Vascular endothelial growth factor receptor 2 (VEGF-R2/KDR), and FGF-2 (Figure 1H).

Figure 1. Vasculogenic EVs fabrication, isolation, and characterization.

A) Schematic describing how engineered vasculogenic EVs are derived from primary human dermal fibroblasts after electroporation with plasmids encoding for the vasculogenic transcription factors genes ETV2, FLI1, and FOXC2 (EFF). Nanoparticle tracking analysis showing B) particle concentration and C) particle size distribution for EFF and control EVs isolated at 12, 24, and 48 hours after electroporation of the donor cells. D) CryoEM images showing single EFF and control EVs isolated 48 hours post-electroporation of the donor cells, showing their intact morphology with a defined lipid bilayer. E) Western blot characterization of EFF and control EVs showing positive expression of the EV marker, CD63 for both EV preparations. F) ETV2, FLI1, and FOXC2 (EFF) upregulation in donor cells at 12, 24, and 48 hours after electroporation compared to sham electroporated cells. G) Dynamics of transcript packing in EFF and control EVs isolated at 12, 24, and 48 hours after electroporation of donor cells. H) Transcript levels of additional vasculogenic factors (VEGF-A, VEGF-KDR, and FGF-2) packed in EFF and control EVs isolated 48 hours after electroporation of donor cells. Data represented as standard deviation Mean+/−SD. # Indicates significant difference with respect to control EVs. * ^# P < 0.05 (n = 3 – 6, One way ANOVA with Fisher LSD multiple comparison test).

Engineered vasculogenic EVs are capable of transfecting HDFs and inducing direct reprogramming towards iECs capable of forming mature vasculature in vivo

EFF EVs collected 48 hours after electroporation of donor cells, were effectively captured and incorporated by recipient cells as early as 6 hours after treatment, with uptake efficiencies of 90.6% and 70.9% for EFF and control EVs, respectively (Figure 2A–B). qRT-PCR analyses confirmed that recipient cells, in this case HDFs, were successfully transfected with the vasculogenic cargo and showed sustained overexpression of EFF for up to 12 hours after treatment with EFF EVs compared to cells treated with control EVs (Figure 2C). Recipient cells showed sustained EFF expression at weeks 1 and 2 post-EV treatment (Figure 2D). We also observed that HDFs treated with EFF EVs were capable of further releasing EFF-loaded EVs for up to 2-weeks after treatment (Figure 2E). Further characterization confirmed the presence of both EFF plasmid DNA and mRNA in these EVs (Supplemental Figure 4A–B). Finding the plasmid DNA in the EVs 1-week after treatment of the HDFs with EFF EVs, suggests that there is a mechanism of EV uptake and subsequent release.

Figure 2. EV uptake by recipient cells and positive transfection of primary HDFs after EFF EV treatment.

A) Representative immunofluorescence micrographs of HDFs treated with fluorescently labeled (red) EFF or control EVs for 6 and 12 hours (cell nuclei were counterstained with DAPI), showing positive uptake of the EVs into subcellular compartments; and B) respective quantification of EV uptake as percentage of positive cells. C) Upregulation of ETV2, FLI1, and FOXC2 (EFF) in recipient cells (HDFs) at 6 and 12 hours after EV treatment. D) Sustained upregulation of ETV2, FLI1, and FOXC2 (EFF) in recipient cells (HDFs) at 1 and 2-weeks after EV treatment. E) Further release of EFF EVs by recipient cells at 1 and 2-weeks after EV treatment. Data represented as Mean+/−SD. * P < 0.05 (n = 3, B-D - Holm-Sidak method and two tailed t-test, E - One way ANOVA with Fisher LSD multiple comparison test).

Additional experiments where the skin of mice was transfected via EFF EVs or via standard bulk electroporation-based transfection were conducted to compare transfection efficiency in vivo. These results demonstrated that in vivo, EFF EVs are more effective in gene transfer compared to standard bulk electroporation-based transfection (Supplemental Figure 5A–B).

Long-term culture experiments also demonstrated the ability of EFF EVs to successfully induce direct reprograming of HDFs into iECs, whereby HDFs treated with EFF EVs showed positive expression of the vascular marker Lectin and the cluster of differentiation 31 (CD31) as early as 1-week after treatment compared to cells treated with control EVs (Figure 3A–B). Positive expression of markers of more mature endothelium such as, von-Willebrand-Factor (VWF), ERG and FGF-2 was observed at 2- and 3-weeks after EV treatment (Figure 3C–G).

Figure 3. EFF EVs induce direct endothelial reprograming of primary HDFs.

A) Successful reprogramming of HDFs towards iECs after treatment with EFF EVs, with positive expression of the endothelia markers CD31 (red) and lectin (green) at 1-week after EV treatment (cell nuclei were counterstained with DAPI), and B) respective quantification of the mean fluorescence intensity for CD31 and Lectin positive cells. C) Immunofluorescence micrographs of iECs showing expression of the more mature endothelial markers VWF (red) at 2 and 3 weeks after EV treatment (cell nuclei were counterstained with DAPI), and D) respective quantification of the mean fluorescence intensity for VWF positive cells. Upregulation of the mature endothelial makers ERG and FGF2 at E) 1-week, F) 2-weeks, and G) 3 weeks post-EV treatment. Data represented as Mean+/− SD. * P < 0.05 (n = 3 – 4, One way ANOVA with Fisher LSD multiple comparison test and Tukey test).

To stablish the ability of EFF EV treated cells to promote new blood vessel formation and increased angiogenesis, 14 days after EFF or control EV treatment, HDFs were embedded in a Matrigel plug and injected in the flank of immunodeficient nude mice (NU/J). (Figure 4A). At day 7 after implantation the animals were euthanized and the Matrigel plugs and surrounding tissues (i.e., skin and muscle) were harvested and cryopreserved for immunohistochemistry (IHC) analysis. Animals implanted with EFF EV-treated cells showed increased angiogenesis with significantly higher vessel area compared to animals implanted with cells treated with control EVs (Figure 4B–C). IHC analysis revealed significantly higher expression of the endothelial markers, VWF and Lectin, in the harvested plugs and the surrounding skin and muscle tissue, for animals implanted with the EFF EV treated cells compared to controls (Figure 4D–E). This data shows the potential of EFF engineered EVs to promote angiogenesis locally, and in the surrounding tissues.

Figure 4. iECs derived from EFF EV treated primary HDFs had the potential to induce formation of mature vasculature in vivo.

A) Schematic diagram of the Matrigel plug assay, where HDFs were embedded in a Matrigel plug at 2-weeks after EV treatment and injected into the flank of immunodeficient nude mice (NU/J) for 1-week. B) Matrigel plugs recovered at 1-week after implantation, showing the potential of iECs to induce formation of mature vasculature in vivo, and C) respective quantification of blood formation in terms of vessel area. D) Representative immunofluorescence micrographs of the Matrigel plugs and surrounding skin and muscle at 1-week post-implantation, showing increased vascularity with upregulation of the endothelial markers VWF (red) and lectin (green) (cell nuclei were counterstained with DAPI), and E) respective quantification of the mean fluorescence intensity for VWF and lectin positive cells. Data represented as Mean+/−SD. * P < 0.05 (n = 5, One way ANOVA with Fisher LSD multiple comparison test).

Engineered vasculogenic EVs promote wound healing

An excisional wound assay was conducted to verify the effect of EFF EVs on wound healing (Figure 5A). EFF or control EVs were topically administered to all animals at day-0 and day-3 post-wounding. Reduction in the wound area was measured at 0, 2, 4, 6, 8, and 10 days after wounding. A significant reduction in wound area was observed as early as 4 days after wounding for animals treated with EFF EVs compared to those treated with control EVs (Figure 5B–C). Animal treated with EFF EVs showed a substantial increase in perfusion by day-10 compared to animals treated with control EVs (Figure 5D–E). Further analysis of skin biopsies at day-10, revealed that animals treated with EEF EVs had improved vascularity with marked expression of the endothelial markers VWF and CD31 (Figure 5F–H). Both EFF EV and control EV treated animals showed positive expression of keratin 5 (KRT5) and keratin 14 (KRT14), which are constitutively expressed in the basal layer of the skin and are involved in proliferation and differentiation of its stratified epithelia⁵¹ (Supplemental Figure 6). Additional experiments were conducted to verify that EVs were successfully internalized by both epidermal and dermal cells (Supplemental Figure 7). However, from previous studies conducted by our group we know that the cells more likely to respond to EFF mediated reprogramming are fibroblasts¹⁸.

Figure 5. EFF EVs promote wound healing in vivo.

A) Schematic illustration of the excisional wound healing assay in nude mice and treatment with EFF or control EVs administrated topical at 0, and 3-days post-wounding. B) Representative images showing accelerated wound closure over time for animals treated with EFF EVs compared to animals treated with control EVs, and C) respective quantification of the percentage of wound closure at 0-, 2-, 4-, 6-, 8-, and 10-days after wounding, showing a significant increase in wound closure by day-4 after wounding for animals treated with EFF EVs. D) Skin perfusion at 10-days post-wounding for animals treated with EFF or control EVs, and E) respective perfusion assessment. F) Representative immunofluorescence micrographs of skin biopsies collected at day-10 post wounding for animals treated with EFF or control EVs, showing positive expression of the endothelial markers vWF (red) and CD31 (yellow), as well as presence of the epithelia marker KRT14 (green), with G) respective zoom-in regions, showing how EFF EV treated animals showed presence of invaginations of the epidermis, resembling hair follicles. H) respective quantification of the mean fluorescence intensity for VWF and CD31 in the wound bed area. Data represented as Mean+/−SD. * P < 0.05 (n = 3, One tailed t-test and One way ANOVA with Fisher LSD multiple comparison test).

DISCUSSION

Our study demonstrates that engineered vasculogenic EVs can be obtained after electroporation of primary HDFs with a cocktail containing expression plasmids encoding for the vasculogenic transcription factor genes, ETV2, FLI1, and FOXC2 (EFF) (Figure 1A–E). We confirmed that these engineered EVs are loaded with plasmid DNA and mRNA encoding for EFF (Supplemental Figure 3). The presence of plasmid DNA content in the EVs has the potential to enhance the number of transcripts produced and packed by donor cells. Moreover, the potential risk of genomic integration is mitigated by the fact that circular plasmids, which are less susceptible to degradation by exonucleases, were used to engineer the donor cells. However, further studies will need to be conducted to evaluate the incidence of plasmid integration, if any^{52, 53}. Overall, this data highlights the ability of engineered EVs to transfer exogenous DNA to target cells.

Additional vasculogenic and angiogenic factors endogenously expressed by the transfected donor cells, such as vascular VEGF-A and -KDR, were successfully packed in the engineered EVs (Figure 1F–H). To the best of our knowledge this is the first report of engineered vasculogenic EVs derived from HDFs. Our results demonstrate that EFF EVs are effectively captured and internalized by recipient cells, and effectively transfect them. A single dose treatment with EFF EVs can promote epigenetic changes in the recipient cells, which induced reprograming of HDFs toward an endothelial phenotype (Figure 3A–D). Moreover, EFF EV treated cells were able to further release vasculogenic EVs for up to 2-weeks after treatment (Figure 2E).

EFF EVs can be incorporated by recipient cells through direct membrane fusion or via the endocytic pathway¹. Endocytosis is divided into two principal sub-types, pinocytosis, and phagocytosis. Recent reports have shown that the pinocytosis pathway is the major route for EVs internalization ^{54, 55}. Although the mechanisms of EVs trafficking within the cell after EV incorporation remains unknown, EVs can effectively escape the endosomal and lysosomal degradative pathways^{56, 57}. Thus, after EV uptake their molecular cargo like plasmid DNA and mRNA can be released into the cytosol without degradation by the lysosomal pathway^{55, 58}. EVs mRNA and plasmid DNA content would be translated in the recipient cells to produce target proteins.

Engineered EVs have the potential to enable novel therapeutic cell-free, non-viral, strategies that can overcome the limitations exhibited by stem cell and progenitor cell-based therapies. Recently, EVs isolated from umbilical mesenchymal stem cells (MSCs) have shown to induce cutaneous wound healing by promoting tissue repair, and angiogenesis^{59 60}. Moreover, EVs derived from MSCs have shown to improve functional recovery, induce neurovascular remodeling, and increase axonal density and synaptophysin-positive area in the ischemic boundary zone using a rat model for middle cerebral artery occlusion (MCAO)⁶¹. However, the EVs isolated from stem cells conserved molecular cues from their donor cells, which may pose additional risks in terms of tumorigenesis. The implementation of human dermal fibroblasts as donor cells circumvents this issue.

Our results indicate that EFF EVs can be used to effectively induce direct reprogramming of adult HDFs to iECs, with positive expression of the endothelial marker CD31 as early as 1-week after EV treatment and sustained expression of VWF for up to 3 weeks after EV treatment (Figure 3B–D). CD31 is a transmembrane protein, which is expressed in early endothelial development⁶², while VWF has heterogeneous levels of expression in vascular tissues (e.g., veins, arteries, arterioles, and capillarity beds), for this reason it is used as an angiogenesis marker⁶³. These epigenetic changes are facilitated by the additional vasculogenic factors packed in the EFF-loaded EVs, such as VEGF-A and -KDR, and FGF2 (Figure 1G–H); as well as those further expressed by the treated cells such as ERG and FGF2 (Figure 3E–G).

Engineered vasculogenic EVs offer a benign method to effectively transfect primary HDFs. Using engineered EVs as nanocarriers to deliver genetic material to target cells can help circumvent limitations imposed by standard electroporation techniques, where the whole cell membrane is exposed to an electric field to achieve transient poration and transfer of the genetic material via electrophoresis. Although effective, this technique can alter the cell membrane composition and integrity, which in turn can affect cell viability, biological structure, and uptake of molecular cargo by target cell⁶⁴. Our results highlight how engineered EV-based nanocarriers offer specific advantages over synthetic and/or viral delivery methods, where EFF EVs are captured with uptake efficiencies of 90.6% and 91.4% at 6- and 12-hours post-EV treatment (Figure 2A–B), they can transfect primary HDFs and induce reprogramming under simple culture conditions (i.e., without the need for differentiation factors). Our findings indicate that human iECs obtained after treatment with EFF EVs can form mature vasculature once transplanted into immunodeficient nude mice (NU/J). Moreover, we observed a robust effect of EFF EVs on wound closure, perfusion, and vascularization at day-10 post-EV treatment. (Figure 4–5 and Supplemental Figure 3).

CONCLUSIONS

EVs have increasingly gained attention due to their capacity to inherently mediate cell-to-cell communication, which makes them a promising candidate for drug and gene delivery applications. Ours study indicates that EFF-loaded EVs can successfully induce direct reprograming of primary HDFs into endothelial cells (iECs). The iECs generated after a single dose treatment with EFF EVs express early endothelial markers such as CD31 and lectin as early as 1-week after treatment, and more mature markers such as VWF as early as 2-weeks post-EV treatment. iECs reprogramed via EFF EV treatment can form mature vasculature in Matrigel plugs transplanted in mice at 1week after EV treatment. In vivo wound healing experiments showed that EFF EVs induced more rapid wound closure compared to animals treated with control EVs. This response can be partially mediated by the sustained release of vasculogenic EVs from recipient cells for up 2-weeks post-EV treatment. These results demonstrate the feasibility of using engineered vasculogenic EVs derived from adult skin fibroblasts, as effective nanocarriers to deliver gene therapies for ischemic disorders.

MATERIALS & METHODS

Cell culture

Primary human dermal fibroblasts (HDF) were isolated from surgical discard skin from healthy female and male patients under IRB approval (2019H028) as previously described⁶⁵. HDF were cultured in DME (Thermo Fisher Scientific) supplemented with 1% penicillin streptomycin, 4% fetal bovine serum (FBS; Gemini BioProducts, West Sacramento, CA), 5 μg/mL bovine insulin (Sigma), 0.1 mM l ascorbic acid-2-phosphate (Sigma), 0.5 μg/mL hydrocortisone (Sigma), and 10 ng/mL epidermal growth factor at 5% CO₂ in a humidified incubator.

DNA plasmid preparation and cell transfection

ETV2, FLI1 and FOXC2 plasmids were purchased from OriGene. All plasmids were expanded via bacterial inoculation. The plasmid purification was made using a DNA ZymoPure II kit (Zymo research, USA. Cat. No. D4201) following the procedure described by the manufacturer. DNA concentrations and quality were assessed using a Nanodrop 2000c Spectrophotemeter (Thermo Fisher Scientific). A cocktail containing plasmids encoding for ETV2, FLI1 and FOXC2 (EFF) were transfected into the donor cells at a 1:1:1 ratio (with 0.05 μg/uL per plasmid) using a Neon^™ Transfection System (Thermo Fisher Scientific). At 1425V, with 1 pulse, for 30ms After transfection, cells were maintained in DME media with 4% of exosome depleted fetal bovine serum (GIBCO cat: A27208–01). The plasmids used in this study included: human-ETV2 (cat: RG213907, OriGene), human-FOX2 (cat: RG223412, OriGene), human-FLI1- (cat: RG200695, OriGene) and the sham vector pCMV6 (cat: PS100010 OriGene) used to produced control EVs.

Extracellular Vesicle isolation

EVs were isolated from culture media at 12-, 24- and 48-hours after transfection of the donor cells. Supernatants were centrifuged at 2000 g for 30 minutes and 4°C to remove dead cells and debris. After centrifugation, total exosome isolation reagent (Thermo Fisher Scientific) was added to the cell-free supernatant and the samples were incubated overnight at 4°C. All samples were subsequently centrifuged at 10,000 g and 4°C for 1 hour and the resulting pelleted EVs were stored at −80°C for subsequent use. EV concentration and size distribution were quantified via nanoparticle tracking analysis using a Nanosight system (Malvern NS300). If needed the samples were diluted to ~10⁷-10⁹ particles/mL to properly fit the detection range for this system⁶⁶. Effective cell transfection of donor and recipient cells were evaluated via qRT-PCR using specific Taqman primers (see Table 1).

Table 1.

Human primers (Taq-Man,ThermoFisher) used to evaluate gene expression analysis.

Gene symbol	Gene name	Gene aliases	Catalog number
GAPDH	glyceraldehyde-3-phosphate dehydrogenase	G3PD, GAPD, HEL-S-162eP	Hs02786624_g1
ETV2	Human ets variant 2	ER71, ETSRP71	Hs00403699_m1
FLI1	Human Friend leukemia virus integration 1	EWSR2, SIC-1	Hs00956709_m1
FOXC2	Human forkhead box C2	FKHL14, LD, MFH-1, MFH1	Hs00270951_s1
ERG	ETS transcription factor	Erg-3, p55	Hs01554629_m1
VEGF-A	vascular endothelial growth factor A	MVCD1, VEGF, VPF	Hs00900055_m1
VEGF-D	c-fos induced growth factor	VEGFD, FIGF	Hs01128657_m1
FGF2	fibroblast growth factor 2	BFGF, FGF-2, FGFB, HBGF-2	Hs00266645_m1
KDR	kinase insert domain receptor	CD309, FLK1, VEGFR, VEGFR2	Hs00911700_m1

Western blot Analysis

Total protein isolation from HDF donor cells, control and EFF EVs was conducted using a lysis buffer containing radioimmunoprecipitation assay buffer (reference no. 89900, Thermo Fisher Scientific), complete protease inhibitor cocktail (reference no. 4693116001, Roche), phosphatase inhibitor cocktail 2 (reference no. P5726–1, Sigma-Aldrich), and 1 mM phenylmethylsulfonyl fluoride (PMSF). All samples were incubated in lysis buffer for 15 min at 4°C and overnight at −20°C; the samples were subsequently vortexed for 15 min, and centrifuged at 10,000 RPM, 4°C for 30 min. The protein concentration for all samples was determined via a standard Bradford protein assay (reference no. 500006, Bio-Rad). 13 ug of total protein was mixed with NuPAGE^™ LDS Sample Buffer (4X) (Thermo Fisher Scientific, cat: NP007) and NuPAGE^™ Sample Reducing Agent (10X) (Thermo Fisher Scientific, cat: NP009). Then, the samples were heated to 85°C for 2 minutes. The protein samples were separated using a 10–20% Tris-glycine gel (Invitrogen, cat: XP10202BOX) and then transferred onto a nitrocellulose membrane (Thermo Fisher Scientific, cat: LC2001). The membrane was subsequently blocked with 5% BSA at room temperature for 1 hour. The antibodies were diluted in TBS-T (1X) with 2.5% BSA and incubated overnight at 4°C. The following antibodies and dilutions were used for this study: β-Actin 1:500 (Cell signaling technology, cat: 8h10d10) and CD63 1:200 (Abcam, cat: ab216130). Lastly, membrane visualization was done using a C-DiGitLI-CORblot scanner.

Gene expression analysis

Total RNA from HDF donor cells, EFF EVs, and control EVs were extracted using Trizol reagent (Thermoscientific cat: 15596018). DNase treatment was performed to remove genomic and plasmid DNA (Thermoscientific cat: 18068015). RNA concentration was measured using a Nanodrop 2000c Spectrophotometer (Thermo Fisher Scientific). Reverse transcription reactions were performed using superscript VILO cDNA synthesis kit (Termo Fisher) with 1000–2500 ng of RNA in 20 μL. ETV2, FLI1, and FOXC2 TaqMan primers were used to evaluate the level of expression via quantitative and absolute real time PCR. Real-time PCR reactions were performed with TaqMan fast advance chemistry (Thermo Scientifc) under the following conditions: 95 °C for 10 min, 40 cycles of 95 °C for 1 min, 60 °C for 1 min, and 72 °C for 1 min using a QuantStudio 3 Real-Time PCR System (Termo Fisher). Human GAPDH (TaqMan, Thermo Fisher. Cat: Hs02786624_g1) was used as a normalizing gene in all qRT-PCR reactions. For the absolute PCRs 10-fold serial dilution of each factor (ETV2, FLI1, and FOXC2) were used as standard curve. Table 1 lists all the primers used in this study.

The plasmid DNA content in the EVs was evaluated via conventional PCR. DNase treatment was performed before the DNA and RNA isolation. Specific primers that target the poly-linker of the engineered plasmid were used VP1.5-Forward (5`-GGACTTTCCAAAATGTCG-3`) and XL39-reverse (3ÀTTAGGACAAGGCTGGTGGG-5`) from OriGene. The following conditions were used to perform the PCRs 1 cycle at 95°C for 1 min, 30 cycles at 95°C for 30 sec, 55°C for 30 sec, and 72°C for 1 min, and 1 cycle at 72°C for 5 min. Products were analyzed by 0.8% agarose gel electrophoresis.

Analysis of EV uptake

Approximately 1.5×10⁵ HDF cells were seeded and allowed to adhere on glass cover slips coated with gelatin (Fisherbrand^™ cat: 1254580). EVs were stained with PKH26 Red Fluorescent Cell Linker (Sigma-Aldrich, cat: PKH26GL-1KT) following the procedure described by the manufacturer. Pellets of EFF or control EVs were resuspended in DME media with 4% of exosome depleted fetal bovine serum (GIBCO cat: A27208–01), and recipient cells were treated with 3.34×10²EVs/cell for 6 and 12 hours.

In vitro reprogramming assays

EFF EVs were used to reprogram HDFs into iEC. For this end, approximately 1.5×10⁵ HDF cells were seeded and allowed to adhere on glass cover slips coated with gelatin (Fisherbrand^™ cat: 1254580). Pellets of EFF or control EVs were resuspended in DME media with 4% of exosome depleted fetal bovine serum (GIBCO cat: A27208–01). Recipient cells were treated with 1.04 ×10⁴ EVs/cell for 48 hours. At 48 hours after EV treatment, the cells were maintained in culture using endothelial cell growth basal medium-2 (EBM; Lonza cat: CC-3156) for 1, 2, and 3 weeks.

Cryo-Electron Microscopy (CryoEM)

The size and morphology of individual EVs was evaluated via Cryo-Electron Microscopy (CryEM). CryoEM grids were frozen using Vitrobot Mark IV system (Thermo Fisher Scientific, Hillsboro). To make the carbon surface hydrophilic, the grids were pretreated using a Pelco EasiGlow system for 30s at 20mA. A small aliquot (3μl) of sample was then applied to the carbon side of the glow-discharged Lacey carbon grid. After blotting away excess liquid, the grid was immediately plunged into liquid ethane to rapidly form a thin film of amorphous ice. The Vitrobot was operated at 4°C and 100% humidity. The blotting force was set at +1. The blotting time was 4s. The frozen grids were clipped into AutoGrids and stored in liquid nitrogen. CryoEM images were captured using a Falcon 3EC direct electron detector under linear mode on a Thermo Scientific^™ Glacios^™ CryoTEM. The microscope was operated at an acceleration voltage of 200kV. The images were collected using EPU software at 57,000x or 6,700x nominal magnification.

Immunocytochemistry

Cells were fixed using 10% formalin for 15 minutes at room temperature, blocked with 5% normal goat serum and incubated with primary antibodies against the endothelial antibody markers VWF (rabbit, ab6994, 1/1000) and CD31 (mouse, ab9498, 1/200), followed by incubation with the secondary antibodies: Alexa 568-tagged α-rabbit (1:200), or Alexa 647-tagged α-mouse (1:200), respectively. Cell nuclei were counterstained with DAPI. Images were captured using a NikonTi2e microscope. Cells incubated without primary antibodies were used as a negative control for staining specificity. The fluorescence intensity was quantified using ImageJ/Fiji and normalized by the area of interest.

Matrigel plug assay

For this assay iECs were used at 14 days after EV treatment. A total of 1×10⁶ iECs were resuspended in 500 uL of Matrigel (ThermoFisher, cat: #3632) and Endothelial Cell Basal Medium-2 (Lonza) without FBS. The mix was injected subcutaneously into the rear flank of 10-week-old male immunodeficient nude mice (NU/J) (Jackson Laboratory, Strain #002019). At 7 days post-implantation the Matrigel plugs, and surrounding tissues (including the skin above and muscle underneath the plugs) were collected, preserved in optimal cutting temperature (OCT) compound (Fisher Healthcare, cat: 4585), and stored at −80°C for subsequent analysis. The Matrigel plugs were sectioned to a thickness of 20 μm using a cryostat. All animal studies performed were approved by the Institutional Animal Care and Use Committee at The Ohio State University (protocol # 016A00000074-R2).

Characterization of transfection efficiency in vivo

C57BL/6 mice (10- to 16-weeks-old, female and male) were obtained from the Jackson Laboratory (Bar Harbor, ME, USA). Mice were depilated 24 hours before the procedure. The mice were anesthetized by continuous 1–3% isoflurane administration throughout the procedure. For in vivo bulk electroporation (BEP), 0.01 ug/uL of the EFF plasmid cocktail was injected intradermally using a U-100 insulin syringe (BD 329424). Then, the injected skin was gently clamped between two electrodes, and a pulse electric field was applied to facilitate electroporation and transfection of the plasmids (200V, 10 ms pulses, 10 pulses). For the EFF EVs, a saline solution containing 1 ×10⁸ EVs/mL was injected intradermally using a U-100 insulin syringe. After 24 hours, a 10 mm skin biopsy was collected. Transfection efficiency was evaluated via qRT-PCRs. pCMV6 plasmid and control EVs were used for comparison purposes for BEP and EFF EVs, respectively. Additionally, uptake of the engineered EVs by skin cells in vivo was confirmed via fluorescence imaging at 24 hours after intradermal injection of fluorescently (red) labeled EFF or control EVs (1 ×10⁸ EVs/mL) using a U-100 insulin syringe. All animal procedures performed in this study were approved by the Institutional Animal Care and Use Committee at The Ohio State University (protocol # 2016A00000074-R2).

Excisional wound healing assay

8-weeks-old male immunodeficient nude mice (NU/J) (Jackson Laboratory, Strain #002019) were used of this assay. Briefly, the dorsal skin of the mice was depilated 24 hours before the procedure. The mice were anesthetized by continuous 1–3% isoflurane administration throughout the duration of the procedure. The back skin was cleaned and sterilized with betaine and 70% ethanol. A sterile 6 mm biopsy punch was used to create a wound on the back skin of each animal. A 0.5 mm thick silicon sheet was placed around the wound, and it was secured with eight interrupted sutures of non-absorbable 6–0 Nylon (Henry Schein, cat: 1016053). After surgery, the mice were treated with EFF or control EVs at a final concentration of 1.8 ×10⁸ EVs/mL in saline solution, and control mouse were treated with saline solution. Finally, a single dose of buprenorphine was administered subcutaneously to control pain. The animals were housed individually after surgery. Skin perfusion was assessed at 10 days after EV treatment, immediately before euthanasia using a high-resolution laser doppler imager (MOORLDI2-HIR, Moor Instruments). All animal procedures performed in this study were approved by the Institutional Animal Care and Use Committee at The Ohio State University (protocol # 2016A00000074-R2).

Immunohistochemistry

Skin biopsies were collected at 10 days post-EV treatment. For this end, 12 mm of skin biopsies were collected, preserved in optimal cutting temperature (OCT) compound (Fisher Healthcare, cat: 4585), and then stored at ~80 °C for subsequent analysis. Preserved skin samples were sectioned and processed for immunohistochemistry staining. Briefly, the samples were fixed with 10% formalin for 15 minutes at room temperature, blocked with 5% normal goat serum and incubated with primary antibodies against the endothelial markers VWF (rabbit, ab6994, 1/1000) and CD31 (mouse, ab9498, 1/200), Keratin 14 (Chicken, 1/200, Biolegend), and Cytokeratin 5 (rabbit, ab52635, 1/200) Skin biopsies collected 24 hours after EFF or control EV injection, were stained using the primary antibody for vimentin (rabbit, ab 92547,1/200) and Keratin 14 (Chicken, 1/200, Biolegend). followed by an incubation with the pertinent secondary antibodies: Alexa 568-tagged α-rabbit (1/200), Alexa 647-tagged α-mouse (1/200), or Alexa 488-tagged α-Chicken (1/200). Cell nuclei were counterstained with DAPI. Skin sections incubated without primary antibodies were used as negative control for staining specificity. All images were captured using a NikonTi2e microscope. For wound the wound healing assay, the fluorescence intensity was quantified in the wound bed using ImageJ/Fiji and all measurements were normalized by the area.

Statistical analysis

Statistical analyses were performed using SigmaPlot 12. Analysis of normality for all data sets were evaluated using a Shapiro-Wilk test. Normal data was evaluated using one tailed t-test, two tailed t-test, and one-way analysis of variance (ANOVA), while non-normally distributed data was analyzed via post hoc Dunńs non-parametric comparison. All data are reported as mean ± standard deviation (SD). The graphics were made using GraphPad Prism7.