Human umbilical vein endothelial cells (HUVECs) and human pericytes underwent VascuNet™ Pericyte Co-Culture Assays (ESI BIO, USA) and were grown in VascuNet Basal Assay Medium supplemented with recombinant human vascular endothelial growth factor (VEGF), epidermal growth factor (EGF), insulin-like growth factor-1 (IGF-1), fibroblast growth factor (FGF) basic, ascorbic acid, heparan sulfate, hydrocortisone hemisuccinate, L-glutamine, 5% fetal bovine serum (FBS), penicillin (100 U/ml), and streptomycin (10 μg/ml) at 37°C in a humidified incubator containing 5% CO₂.

Subcutaneous white adipose tissue (scWAT) and visceral gonadal WAT (gWAT) white fat pads were dissected and processed for cell isolation. Mouse adipocytes were obtained according to a previously reported method with minor modifications (24). Mouse WAT fat pads were dissected and gently minced with scissors for 2–5 min or until no visible tissue fragments remained. The minced tissue was digested in a shaking water bath at 37°C and 170 rpm with 10 ml of phosphate-buffered saline (PBS) containing 1% bovine serum albumin (BSA), 400 nM of adenosine, 50 μg/ml of gentamicin, 2.5 mg/ml of dispase II, and 10 mM of CaCl₂ in the presence of collagenase D (Roche). Minced WAT was digested using collagenase, then the adipocytes were washed three times with 40 ml wash buffer [1% bovine serum albumin (BSA), 400 nM adenosine, 50 μg/ml gentamicin in PBS, pH 7.4] at room temperature. The infranatant was removed using a syringe and needle after allowing the cells to float for more than 5 min after each round of washing. After removing the wash buffer from the last round, isolated adipocytes were filtered using a 100- or 300-μm nylon mesh strainer and placed in Dulbecco's PBS (DPBS) for staining. Primary adipocytes were not centrifuged, since this would result in greater cell breakage and sample detritus. When fixation was required, adipocytes were fixed with fixative buffer (2% paraformaldehyde and 1% sucrose in DPBS) for 30 min with gentle agitation. The cells were resuspended in the medium after washing with wash buffer (24).

We applied three markers to flow cytometry (FCM): CD146-fluorescein isothiocyanate (FITC), nestin-phycoerythrin (PE), and NG2-PE to assess and sort the pericyte component of cultured mouse WAT adipocytes. Cells were collected from culture dishes using trypsin ethylenediaminetetraacetic acid (EDTA), fixed in 4% paraformaldehyde at pH 7.4, and dissolved in PBS. For FCM, a single cell was treated with the appropriate primary antibodies for 30 min and then rinsed with PBS supplemented with 1% FBS. After washing twice to remove the primary antibody, samples were re-incubated with an FITC- or PE-conjugated secondary antibody (Invitrogen, USA) at a dilution of 1:500 for 20 min. Prior to initiating FCM, cells were washed twice more with wash buffer. FCM was performed using FACSCalibur (Beckman Coulter, California, USA) (25).

A Matrigel matrix (Corning, USA) was thawed on a 24-well plate. After 1 h at ambient temperature, the coated 24-well plate was placed in an incubator at 37°C with a humidified environment of 5% CO₂ for 1–2 h prior to plating. Each well contained 300 μl of the human umbilical vein endothelial cells (HUVECs) and pericytes (10:1 ratio) cell suspension in the VascuNet Basal Assay Medium (ESI BIO, USA) (26).

Cell migration was determined using a wound healing test. Clear lines were formed using a sterile 200 μl pipette tip after cells were planted at the same density per well for 12 and 24 h. The cells were then grown continuously for 12–24 h in media supplemented with 5% FBS. At 0, 12, and 24 h following initiation of the wound, photomicrographs with the scratches were taken at prefixed positions. ImageJ was used to determine the migration distance. Each experiment was conducted a minimum of three times. In each scenario, the mean was calculated as the average of the experimental repetitions (27).

For 5 days, mice were intraperitoneally injected with a single high dose of 50 mg/kg of streptozotocin (STZ). Shock fatalities were reduced by substituting 5% sucrose water for drinking water. The tail was pricked with a lancet to draw blood. After 2 weeks of STZ, body weight and blood glucose levels were examined daily until a diabetic condition was confirmed. Diabetic mice were defined as those with a glucose level of more than 300 mg/dl. Mice that had a successful course following diabetes induction were then assigned to investigations.

All mice were kept in compliance with the National Institutes of Health (NIH) and institution-approved animal care protocols at the CHA University Veterinary Service Center. The CHA Administrative Panel on Laboratory Animal Care authorized all operations. Male mice aged 8–10 weeks (Orient Bio, Korea) were divided into four groups: “untreated control” (control), “pericyte injection” (pericyte), “diabetes mice without pericyte injection” (DM without pericyte), and “diabetic mice with pericyte injection” (DM with pericyte) (n = 10 wounds per treatment group). Under anesthesia, the dorsal surface of the mouse was shaved. On either side of the midline, two full-thickness incisions 4 mm in diameter were produced. To avoid wound contraction, the wounds were circumscribed with donut-shaped silicone splints. A 10-μl mixture was pipetted over a type I collagen sponge and incubated for 1 day to completely remove the sponge (Advanced BioMatrix, California, USA). All the wounds were then treated with a commercially available clear film covering 3M Tegaderm Film (Minnesota, USA). Photographs were taken on each day and the wound area was measured using ImageJ software (NIH, Maryland, USA). At different time points, the percentage of the original wound area was calculated with reference to the wound area on the day of surgery (28).

A total of 14 days after wounding, wound beds and underlying muscle were removed, fixed for 4 h in 10% formalin (Sigma-Aldrich), treated with 100% ethanol and xylenes, and embedded in paraffin. Serial sectioning of 10 wounds (7 μm) perpendicular to the wound surface was performed in the rostral-to-caudal direction for each group. For wound analysis, every tenth segment of the whole wound bed was stained with H&E and Masson's trichrome. Immunohistochemical analysis determining the blood vessel formation capacity of the pericytes was conducted according to the manufacturer's protocol using a Ready-to-use IHC/ICC kit (BioVision Incorporation). Tissues were preserved for 24 h at room temperature with 4% formaldehyde. Tissues fixed in paraffin were sectioned into 5 μm thick slices, deparaffinized, rehydrated, and subjected to antigen retrieval. After blocking endogenous peroxidase activity with 3% hydrogen peroxide (H₂O₂), the slides were treated at room temperature with anti-Ki67 (Abcam, ab15580) and anti-CD31 (Abcam, ab28364) antibodies. Additionally, angiogenesis was identified by using immunohistochemistry with antibodies to α-smooth muscle actin (α-SMA), collagen 1 (Col 1), and Nestin. Following incubation with horseradish peroxidase (HRP)-anti-mouse immunoglobulin G (IgG) polymer and washing with PBS, the tissue sections were treated at room temperature with 3,3′-diaminobenzidine. The images were taken with an Eclipse Ts2 microscope at 100 × magnification.

GraphPad Prism program version 6 (GraphPad Software Incorporation, San Diego, California, USA) and R software were used to conduct all the statistical analyses (version 3.2.4; R Foundation for Statistical Computing). Additionally, when only two groups were assessed, the Mann–Whitney U test was utilized to determine significance. The significance level was set at p ≤ 0.05 (^**p ≤ 0.01, ^****p ≤ 0.0001). Unless otherwise specified, all in-vitro studies were repeated at least three times.

We used three-marker polychromatic FCM to assess and sort the pericyte component of cultured mouse WAT adipocytes. Figure 1 shows the expression degree of the three pericyte markers. The frequencies of pericyte markers in the non-diabetic and diabetic conditions were as follows: CD146-FITC, 53.22 and 42.82%; Nestin-PE, 47.69 and 22.38%; and NG2-PE, 40.75 and 18.25%, respectively (Figure 1A). FCM results show decreased levels of the three indicators in diabetic mice compared to those in nondiabetic mice. The difference in expression levels of these three markers is shown in the overlay histogram (Figure 1B).

At 5 mM glucose concentrations, the capillary formation capacities of the HUVEC group and the HUVEC-with-pericytes group were similar, as were numerical data such as number of pieces, number of branches, and total length. However, the results for 30 mM glucose concentration demonstrated a substantial change, involving thinning of the capillaries and a decrease in the number of branches compared to the observations for the 5 mM glucose concentration (Figure 2A). At a 30 mM glucose concentration, the number of pieces and branches increased significantly in the HUVEC-with-pericytes group compared to that in the HUVEC group (Figure 2B).

We evaluated the contribution of morphological differences induced by pericytes to cell migration using a wound healing assay and time-lapse microscopy. At 5 mM glucose concentrations, the HUVEC-with-pericytes group showed faster wound healing, almost completing the closure by 24 h, whereas the HUVEC group progressed slowly. However, at 30 mM glucose conditions, slow, almost negligible migrations were seen in the HUVEC group, which showed significant progress with the addition of pericytes (Figure 2C). Analysis of wound closure was performed by a semi-quantitative approach, using the breadth of the wounds to estimate the relative wound closure. At a 30 mM glucose concentration, pericyte-treated cells migrated at a remarkably faster rate than that of the HUVECs only in the initial time period (Figure 2D).

The effects of the pericyte treatments were assessed in vivo using a diabetic mouse model. The development of diabetes by STZ treatment was verified by blood glucose measurements (Figures 3A,B). Figure 3C shows the excisional wounds on 0, 2, 7, and 14 days after operation in the four groups. On day 2, it was confirmed that the wound size decreased in the DM-with-pericyte group compared to the DM-without-pericyte group. On day 14, the wound area of the DM-with-pericyte group was 0% and that of the DM-without-pericyte group was 36%, showing a statistically significant difference (Figures 3D,E).

Blood vessels create continuous networks parallel to the skin's surface, which can be seen by techniques used to assess blood vessel health, such as H&E and Masson's trichrome staining of dermal scars. The DM-without-pericyte group showed incomplete blood vessel formation and no regeneration of the dermal collagen layer. However, the other three groups—control, pericyte, and DM-with-pericyte—were able to repair their collagen layers successfully (Figures 4A,B). CD31 and Ki67 co-staining were used to assess angiogenesis. Both the markers were remarkably increased in the DM-with-pericyte group compared to those in the DM-without-pericyte group (Figure 4C). In addition, immunohistochemical analysis showed that α-SMA, Col 1, and Nestin expression were higher in the DM-with-pericyte group compared to those in the DM-without-pericyte group, indicating that angiogenesis was induced in these groups (Figure 4D).