Embryonic stem cell-derived extracellular vesicle-mimetic nanovesicles rescue erectile function by enhancing penile neurovascular regeneration in the streptozotocin-induced diabetic mouse

Extracellular vesicles (EVs) have attracted particular interest in various fields of biology and medicine. However, one of the major hurdles in the clinical application of EV-based therapy is their low production yield. We recently developed cell-derived EV-mimetic nanovesicles (NVs) by extruding cells serially through filters with diminishing pore sizes (10, 5, and 1 μm). Here, we demonstrate in diabetic mice that embryonic stem cell (ESC)-derived EV-mimetic NVs (ESC-NVs) completely restore erectile function (~96% of control values) through enhanced penile angiogenesis and neural regeneration in vivo, whereas ESC partially restores erectile function (~77% of control values). ESC-NVs promoted tube formation in primary cultured mouse cavernous endothelial cells and pericytes under high-glucose condition in vitro; and accelerated microvascular and neurite sprouting from aortic ring and major pelvic ganglion under high-glucose condition ex vivo, respectively. ESC-NVs enhanced the expression of angiogenic and neurotrophic factors (hepatocyte growth factor, angiopoietin-1, nerve growth factor, and neurotrophin-3), and activated cell survival and proliferative factors (Akt and ERK). Therefore, it will be a better strategy to use ESC-NVs than ESCs in patients with erectile dysfunction refractory to pharmacotherapy, although it remains to be solved for future clinical application of ESC.

Metabolic variables. Body weight was significantly lower in the diabetic mice than that of control mice.
In addition, diabetic group exhibited significantly increased fasting and postprandial blood glucose concentrations compared with the controls. No significant differences in body weight and blood glucose levels were found between the diabetic groups, regardless of the treatment given (Tables 1 and 2). ESC-NVs fully, and ESCs partially restore erectile function in the diabetic mice. To determine the physiological relevance of intracavernous injection of ESC or ESC-NVs, we performed nerve-induced erectile function study. Erectile function parameters, such as the ratios of maximal intracavernous pressure (ICP) and total ICP to mean systolic blood pressure (MSBP), were profoundly decreased in phosphate-buffered saline (PBS)-or HEPES (4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid)-buffered saline (HBS)-treated diabetic mice compared with age-matched controls (Figs. 2 and 3). Dose-dependent experiments revealed that ESC at a concentration of 3 × 10 5 cells/20 µL partially restored erectile function, which reached up to 78% (maximal ICP) or 76% (total ICP) of control values (Fig. 2). In contrast, intracavernous injection of ESC-NVs restored erectile function in a dose dependent manner in diabetic mice, showing almost complete recovery of erection parameters at a concentration of 1.0 µg/20 µL, which reached up to 95% (maximal ICP) or 97% (total ICP) of normal value (Fig. 3). ESC-NVs (1.0 µg/20 µL) demonstrated superior erectile function recovery than ESC (3 × 10 5 cells/20 µL) (see Supplemental Fig. 1). No detectable differences in MSBP were found among the experimental groups (Tables 1 and 2). ESC-NVs promote angiogenesis under diabetic conditions. We performed immunofluorescent staining with antibodies against smooth muscle α-actin, platelet/endothelial adhesion molecule 1 (PECAM-1), and neuron-glial antigen 2 (NG2) in the cavernous tissue of control and diabetic mice 2 weeks after treatment of ESC-NVs (1.0 µg/20 µL). The cavernous expression of smooth muscle cell, endothelial cell, and pericyte contents was significantly lower in the HBS-treated diabetic mice than in the control mice. Intracavernous injection of ESC-NVs induced complete restoration of smooth muscle cell and endothelial cell contents, and partial restoration of pericyte content in the diabetic mice (Fig. 4).
An in vitro matrigel assay revealed impairments in tube formation in MCEC or MCP exposed to high-glucose condition, and these impairments were partially restored by treatment with ESC-NVs (Fig. 5a-c). An ex vivo aortic ring assay revealed significantly decreases in the average length and branch number of outgrowing microvessels in aortic segments exposed to high-glucose condition compared with that in the segments exposed www.nature.com/scientificreports www.nature.com/scientificreports/ to normal-glucose condition. ESC-NVs significantly enhanced the outgrowth of microvessels from aortic rings under high-glucose condition (Fig. 5d,e).
ESC-NVs induce neural regeneration under diabetic conditions. The expression of βIII tubulin in corpus cavernosum was significantly lower in the HBS-treated diabetic mice than in the control mice, whereas the neuronal cell content was completely restored by treatment with ESC-NVs (Fig. 6a,c). ESC-NVs also profoundly enhanced neurite sprouting in an ex vivo cultured MPG tissue exposed to high-glucose condition (Fig. 6b,d).
ESC-NVs increase the expression of angiogenic and neurotrophic factors, and enhance cell proliferative and survival pathway. Western blot analysis showed that the cavernous expression of hepatocyte growth factor (HGF) and angiopoietin-1 (Ang1) protein was significantly lower and angiopoietin-2 (Ang2) protein expression was significantly higher in the HBS-treated diabetic mice than in the age-matched controls. The expressions of these angiogenic factors were returned to control values after treatment with ESC-NVs (Fig. 7a,d-f).  www.nature.com/scientificreports www.nature.com/scientificreports/ To test whether the effects of ESC-NVs was mediated by the production of neurotrophic factors, we performed Western blot analysis for nerve growth factor (NGF) and neurotrophin-3 (NT-3). The cavernous expression of NGF and NT-3 was significantly lower in the HBS-treated diabetic mice than in the control mice. ESC-NVs restored the cavernous expression of NGF and NT-3 in the diabetic mice ( Fig. 7b,g,h).
Moreover, ESC-NVs also induced phosphorylation of ERK, Akt, and endothelial nitric oxide synthase (eNOS) in the corpus cavernosum of diabetic mice (Fig. 7c,i-k).

Discussion
Here, we examined the efficacy of ESC-NVs in a mouse model of diabetic ED. Intracavernous administration of ESC-NVs induced almost complete recovery of erectile function in the diabetic mice, whereas intracavernous injection of ESC partially restored erectile function. The beneficial effects of ESC-NVs were accomplished by restoring cavernous contents of endothelial cells, smooth muscle cells, pericytes, and neuronal cells in the diabetic mice in vivo; by promoting tube formation in primary cultured MCEC and MCP under high-glucose condition in vitro; and by accelerating microvascular and neurite sprouting from MPG under high-glucose condition ex vivo, respectively. ESC-NVs induced the expression of angiogenic and neurotrophic factors, and activated cell survival and proliferative factors in the diabetic mice in vivo.
Although the EVs do not have a potential of malignant transformation, their proangiogenic and proliferative effects may accelerate cancer progression. Moreover, previous study reported that low dose of endothelial cell-derived EVs increased tube formation, whereas a high concentration had an inhibitory effect 20 . Therefore, it is particularly important to determine optimal dose of EVs to minimize side effects, while enhancing their therapeutic efficacy. In the present study, we determined optimal dosage of ESC-NVs (0.1 μg, 0.5 μg, 1 μg, 2 μg, or 5 μg in 20 μL of HBS, respectively) and obtained a maximal erectile function recovery at a concentration of 1 μg/20 μL.
It was demonstrated that EVs released by endothelial cells contain β1 integrin and matrix metalloproteinase-2 and −9, and promoted endothelial cell invasion and capillary-like tube formation 21 . In the present study, ESC-NVs increased the cavernous expression of Ang1 and HGF, and decreased the expression of Ang2 in diabetic mice. Ang1 is a secreted protein ligand for tyrosine kinase with immunoglobulin and epidermal growth factor homology domain-2 (Tie2, also called Tek). Ang1 has a major role in blood vessel remodeling, maturation, and stabilization 22,23 . Ang2 is an endogenous antagonist of Ang1 24 . We recently reported in mouse models of type I and type II diabetic ED that intracavernous administration of synthetic Ang1 restores erectile function by enhancing endothelial cell regeneration 25,26 . HGF is also known as a potent angiogenic factor that induces migration of endothelial cells and pericytes 27 . We found that recombinant human-HGF protein completely restored cavernous endothelial cell and pericyte contents, and decreased oxidized LDL leakage in the diabetic mice 3,27 . Therefore, the regulation of these angiogenic factors may be an important mechanism by which ESC-NVs enhance angiogenesis under diabetic conditions. www.nature.com/scientificreports www.nature.com/scientificreports/ After peripheral nerve injury, Schwann cells are known to release cytokines and secrete neurotrophic factors that guide neural regeneration 28,29 . It was reported that Schwann cell NVs, but not fibroblast NVs, specifically enhanced neurite growth in vitro 15 . In this regard, the induction of neurotrophic factors (NGF and NT-3) and subsequent neurite sprouting from ex vivo cultured MPG under high-glucose condition as well as axonal regeneration in the diabetic mice in vivo by ESC-NVs are noteworthy. However, it remains to clarify the sources of growth factors whether these neurotrophic factors as well as angiogenic factors are directly derived from ESC-NV cargo, or endogenously synthesized secondarily from penile neurovascular regeneration. www.nature.com/scientificreports www.nature.com/scientificreports/ Akt is a serine/threonine kinase and downstream signaling mediator of phosphatidylinositol 3-kinase (PI3K), and the activation of PI3K/Akt pathway is known to enhance survival of the various cell types 30 . Activation of ERK pathway is reported to enhance cell proliferation 31,32 . EVs generated in response to interleukin-3 stimulation are known to increase ERK activation and cyclin D1 transcription, and to promote angiogenesis 33 . Moreover, endothelial colony-forming cell-derived EVs enhanced neovascularization and promoted cutaneous wound healing in diabetic rats by activating ERK signaling in endothelial cells and by stimulating the expression of angiogenic molecules 34,35 . EVs isolated from Akt-overexpressing mesenchymal stem cells are also known to stimulate endothelial cell migration, proliferation, and tube-like formation in vitro 36 , and to increase the formation of blood vessel in vivo. In the present study, ESC-NVs enhanced phosphorylation in endothelial cell survival through the PI3K/Akt pathway 23 and to induce endothelial cell proliferation by activation of the ERK pathway 23,31,32 . From these findings, we believe that ESC-NV-mediated increase in angiogenic factors in the corpus cavernosum of diabetic mice may play a key role in activating signaling pathway involved in cell survival and proliferation.
We used the most potent stem cell source, ESC, to isolate EV-mimetic NVs and to confirm the efficacy of ESC-NVs in diabetic ED, although ethical concerns might be raised when those are applicable in clinical situation. The cellular or organ source of EVs is also reported to be of great importance, as shown by the in vivo tracking study which demonstrated that intravenously administered EVs derived from kidney embryonic cells are taken up mainly by the kidney 37 . Thus, it will be intriguing to compare the results of this study with those of future studies using EVs or EV-mimetic NVs derived from a variety of cells, such as endothelial cells, smooth muscle cells, or pericytes isolated from orthotopic organ, i.e., erectile tissue.
The functionality of EVs is strongly influenced by the microenvironment 38 or cytokine stimulation 33,39 . For example, hypoxia stimulation and preconditioning of stem cells with platelet-derived growth factor or an endothelial differentiation medium favored the release of EVs with vasculogenic potential and enhanced their proangiogenic activity 33,38,39 . Therefore, it will be valuable to evaluate whether the several stimuli or modification of culture conditions would result in better outcomes.
Our study has some limitations. We did not screen for the development of ED before treatment of ESC-NVs because of invasive nature of the nerve-induced erectile function study. We demonstrated a short-term efficacy of ESC-NVs in a mouse model of diabetic ED. Further studies are needed to test whether ESC-NVs would induce durable erectile function recovery in a variety of animal models for ED.  www.nature.com/scientificreports www.nature.com/scientificreports/ conclusions Our study demonstrates a unique function of ESC-NVs in the diabetic ED. ESC-NVs ameliorates erectile function in diabetic mice by enhancing penile neurovascular regeneration and demonstrates superior effects than ESC. Local treatment with EV-mimetic NVs may represent a promising therapeutic strategy for the treatment of ED caused by vascular and neural diseases.  www.nature.com/scientificreports www.nature.com/scientificreports/ nonessential amino acids (Gibco), and 0.1 mM β-mercaptoethanol (Gibco) at 37 °C/5% CO 2 . Media was changed daily, and cells were passaged every 2 to 3 days.

Preparation of ESC-NVs.
ESC-NVs were prepared as described previously 19 . Briefly, mouse ESC were rinsed with PBS. Adherent cells were detached using 0.25% Trypsin-EDTA (Invitrogen) and re-suspended in HEPES buffer solution (Gibco). ESC-NVs were produced using a mini extruder system (Avanti Polar Lipids, Birmingham, AL, USA). Cell suspension was sequentially extruded using 10, 5, and 1 μm pore-sized polycarbonate membrane (Nuclepore, Whatman Inc., Clifton, NJ, USA) ten times across each filters. To form a step gradient, 50% iodixanol (1 mL; Axis-Shield PoC AS, Oslo, Norway) was placed at the bottom of an ultracentrifuge tube, overlaid with 10% iodixanol (2 mL) and the extruded samples (7 mL), and then ultracentrifuged at 100,000 g for 2 hours at 4 °C. The interface layers between the 10% and 50% iodixanol were further pelleted at 100,000 g for 2 hours at 4 °C. NVs were filtered with 0.45 μm filter and stored at −80 °C until use (Fig. 1a).
Transmission electron microscopy. The purified ESC-NVs were applied to glow-discharged carbon-coated copper grids (Electron Microscopy Sciences, Fort Washington, PA, USA). After ESC-NVs had been allowed to be absorbed onto the grid for 1 hour, the grids were fixed with 4% paraformaldehyde for 10 minutes and rinsed with droplets of deionized water and then, negatively stained with 2% uranyl acetate (Ted Pella, Redding, CA, USA). Electron micrographs were recorded with a JEM 1011 microscope (JEOL, Tokyo, Japan) at an acceleration voltage of 100 kV as described previously 19 .
Dynamic light scattering. The size distribution of ESC-NVs was measured with Zetasizer Nano ZS (Malvern Instrument Ltd., Malvern, U.K.). The size distribution based on relative abundance was determined by an infrared light (wavelength = 633 nm) passing through the sample at the scattered intensity of 10× for 30 seconds as described previously 19 .

Animals and treatments.
Eight-week-old male C57BL/6 mice were used in this study. The experiments were approved by the Institutional Animal Care and Use Committee of Inha University (Assurance Number: INHA 171129-527) and performed in accordance with relevant guidelines and regulations. Diabetes was induced by intraperitoneal injection of multiple low doses of streptozotocin (STZ, 50 mg/kg body weight in 0.1 M citrate buffer, pH 4.5) for 5 consecutive days as described previously 40 . Animals were considered diabetic if their nonfasting glucose levels were greater than 300 mg/dL. Eight weeks after diabetes was induced, the mice were anesthetized with intramuscular injections of ketamine (100 mg/kg) and xylazine (5 mg/kg) and placed supine on a thermoregulated surgical table.
To test the efficacy of ESC-NVs, the mice were distributed into seven groups (N = 5 per group): age-matched controls and STZ-induced diabetic mice receiving repeated intracavernous injections of HBS (days −3 and 0; 20 μL) or ESC-NVs (days −3 and 0; 0.1 μg, 0.5 μg, 1 μg, 2 μg, or 5 μg in 20 μL of PBS, respectively). ESC-NVs were given twice, because our pilot experiments demonstrated that a single intracavernous injection of ESC-NVs resulted in a partial recovery of erectile function (data not shown). To minimize leakage of the ESC or ESC-NVs into systemic circulation, blood drainage via the dorsal veins was halted by circumferential compression of the penis at the base with an elastic band immediately before injection, and the compression was released at 30 minutes after the injection 41 . We evaluated erectile function by cavernous nerve electrical stimulation 2 weeks after treatment. A separate group of animals was used for histologic examination and biochemical study.
Measurement of erectile function. The mice from each group were anesthetized with ketamine (100 mg/ kg) and xylazine (5 mg/kg) intramuscularly. Erectile function was measured as described previously 40 . Briefly, the bladder and prostate were exposed through a midline abdominal incision. The MPG and cavernous nerve were identified posterolaterally to the prostate on one side, and bipolar platinum wire electrodes were placed around the cavernous nerve for electrical stimulation. The penis was denuded of skin, and a 26-gauge needle filled with 250 U/mL of heparin was inserted into one side of the corpus cavernosum for monitoring ICP with a Statham P23 pressure transducer connected to a computerized system for data acquisition (Biopac Systems, Goleta, CA, USA), which was interfaced to a personal computer for recording and data analysis. Stimulation parameters were 5 V at a frequency of 12 Hz, a pulse width of 1 ms, and a duration of 1 minute. During tumescence, the maximal ICP was recorded. The total ICP was determined by the area under the curve from the beginning of cavernous nerve stimulation to a point 20 seconds after stimulus termination. Systemic blood pressure was measured by using a noninvasive tail-cuff system (Visitech systems, Apex, NC, USA). The ratios of maximal ICP and total ICP (area under the curve) to MSBP were calculated to adjust for variations in systemic blood pressure.
Tube formation assay. The MCEC and MCP were prepared and maintained as described previously 3 . The tube formation assay was performed to assess the angiogenic capacity of ESC-NVs in MCEC or MCP. About 100 µL of growth factor-reduced matrigel was dispensed into 96-well tissue culture plates at 4 °C. After gelling at 37  www.nature.com/scientificreports www.nature.com/scientificreports/ DMEM medium. The assay was performed in a CO 2 incubator and the plates were incubated at 37 °C for 24 hours. Images were obtained with a phase-contrast microscope and the numbers of tubes in each well of the plate were counted at a screen magnification of ×40 (N = 4 per group). Only integrated tubes were counted.
Aortic ring assay. Aortic ring assay were performed as described previously 42 . Aortas were harvested from 8-week-old C57BL/6 mice (N = 4 per group). The aortic rings were placed in the 8-well Nunc Lab-Tek Chamber Slide System (Sigma-Aldrich, Saint Louis, MO, USA) and sealed in place with an overlay of 50 μL matrigel. The aortic rings were cultured in medium 199 with 20 ng/mL of basic fibroblast growth factor and 1% penicillin/ streptomycin for 5 days. The aortic segments and sprouting cells were fixed in 4% paraformaldehyde for at least 30 minutes and used for immunofluorescent staining.
Ex Vivo neurite sprouting assay. The mouse major pelvic ganglion tissues were prepared and maintained as described previously 43 with minor modifications. The MPG tissues were isolated from male mice using a microscope, transferred into sterile vials containing Hank's balanced salt solution (Gibco), and then rinsed and washed twice in PBS. The MPG tissues were cut into small pieces and the samples plated on poly-D-lysine hydrobromide-coated (Sigma-Aldrich) 12-well plate. The whole MPG tissue was covered with matrigel and the culture plate placed on ice for 5 minutes prior to incubation at 37 °C for 10-15 minutes in a 5% CO 2 atmosphere. We added 1 mL of complete Neurobasal medium (Gibco) supplemented with 2% serum-free B-27 (Gibco) and 0.5 nM GlutaMAX ™ -I (Gibco). The dishes were then incubated at 37 °C in a 5% CO 2 atmosphere. Three days after incubation, we evaluated neurite outgrowth. establishment of In Vitro or Ex Vivo experimental systems that mimic diabetic ED. To mimic an in vivo or ex vivo condition for diabetes-induced angiopathy and neuropathy, primary cultured cells or tissues were serum-starved for 24 hours and then exposed to normal-glucose (5 mmol) or high-glucose (30 mmol; Sigma-Aldrich) conditions for 2 days (MCEC and MCP), 3 days (MPG tissue), or 5 days (aortic ring) as described previously 42 .
Histological examinations. The penis tissue (N = 6 per group) and cultured MPG (N = 4 per group) were fixed in 4% paraformaldehyde for 24 hours at 4 °C as described previously 42 . Frozen tissue sections (12-μm or 60-μm thick) were incubated with antibodies to smooth muscle α-actin (Sigma-Aldrich; 1:50), NG2 (Millipore, Temecula, CA, USA; 1:50), PECAM-1 (Millipore; 1:50), or βIII tubulin (Abcam, Cambridge, UK; 1:50) at 4 °C overnight. After several washes with PBS, the tissues were incubated with tetramethyl rhodamine isothiocyanateor fluorescein isothiocyanate-conjugated secondary antibodies (Zymed Laboratories, South San Francisco, CA, USA) for 2 hours at room temperature. Samples were mounted. Signals were visualized and digital images were obtained with a confocal microscope (FV1000, Olympus, Tokyo, Japan). Quantitative analysis of histologic examinations was done with an image analyzer system (National Institutes of Health [NIH] Image J 1.34, http://rsbweb. nih.gov/ij/) and we analyzed the histologic data in a blinded manner. Statistical analysis. The results are expressed as mean ± SE. For parametric data, intergroup comparisons were made by one-way ANOVA followed by Newman-Keuls posthoc tests. We used the Kruskal-Wallis test to compare nonparametric data. Probability values less than 5% were considered significant. We used SigmaStat 3.11 software (Systat Software) for statistical analyses. We also performed a power analysis to determine the minimum number of animals that need to be included in the present study (https://clincalc.com/stats/samplesize.aspx).