Seventy male Sprague–Dawley (SD) rats aged 5 weeks and weighing 150–160 g were purchased from Hubei Research Center of Laboratory Animals (SCXK [E] 2015-0018). The rats were housed in a certified animal care facility, and housing conditions (5 animals per cage, 22 ± 2°C, 55 ± 5% relative humidity, 12 h artificial light/dark cycle) with ad libitum access to water and regular Rat chow (Huafukang, Beijing, CN) for 1 week to adapt to the laboratory environment. Food intake and body weight were measured on a weekly basis.

T2DM was induced using a combination of 2-week high-fat diet (Research Diet D12492 60 kcal% fat, USA), followed by a low dose (35 mg/kg) of streptozotocin (STZ, Sigma Chemical Co., St. Louis, MO) intraperitoneal injection and continued a high-fat diet for another 2 weeks (19). The fasting blood glucose was measured before STZ injection and 2 weeks after STZ injection by using a glucometer (Johnson & Johnson Co., New Brunswick, NJ, USA). The T2DM rat model was successfully established when the fasting blood glucose value of more than 16.7 mmol/L was measured at 2 weeks after STZ injection.

Bone marrow was harvested from 4-week-old male SD rats [Hubei Research Center of Laboratory Animals, SCXK [E] 2015-0018, Wuhan, China] after flushing with phosphate-buffered saline (PBS). The suspension was added into lymphoprep (Axis-Shield, Oslo, Norway) and centrifuged at 400 g for 30 min. Immed iately after isolation, mononuclear cells were washed twice with PBS and seeded on fibronectin (Sigma-Aldrich, St. Louis, MO)-coated 25-cm² culture flasks (Corning Inc., NY) at a density of 2 × 10⁷ cells/flask, which cultured with EGM-2MV Bullet Kit medium (Lonza, MD, USA) at 37°C under 5% CO₂. We changed the culture medium every 3 days. We identified EPCs by their property of uptaking acetylated LDL (ac-LDL) and Ulex europaeus agglutinin-1 (UEA-1). Cells with double positive staining were identified as differentiating EPCs (18).

The lentivirus vectors used in our study were procured from the commercial sources (Genechem Co., Ltd., Shanghai, China). Lentivirus constructs that express the GFP and APN gene to get the LV-GFP-APN vector, the vector without insertion of APN gene, was used as control (LV-GFP). Virus suspensions were stored at −80°C until transfection.

After being cultured for 7 days, EPCs were harvested and divided into three parts. One-third of the EPCs continued to be cultured (EPCs), one-third of the EPCs were transfected with LV-GFP-APN (LV-APN-EPCs), and one-third of the EPCs were transfected with LV-GFP (LV-EPCs). LV-APN-EPCs and LV-EPCs were seeded in 25-cm² culture flasks (1 × 10⁵ cells/flask) and preincubated for 12 h with 4 ml of the EGM-2MV medium at 37°C under 5% CO₂, respectively. After 12 h of culture, the medium was removed, and 2.5 ml of medium without FBS was replaced in the flasks. 5 × 10⁸ LV-GFP-APN with 20 μl and 5 μg/ml polybrene were added into the cells (LV-APN-EPCs) and 1 × 10⁹ LV-GFP with 10 μl and 5 μg/ml polybrene were added into the cells (LV-EPCs) and incubated for 12 h. Then, 2 ml of PBS was used to wash the cells twice to remove any free lentivirus, and 4 ml of fresh medium was added to the flasks, cultured at 37°C under 5% CO₂. After 72 h of culture, we visualized enhanced green fluorescent protein (EGFP) by fluorescence microscopy to evaluate the results of the lentivirus transfection (18). We measured APN expression from EPCs after gene transduction. Expression of APN was detected by Western blot analysis (18).

Seventy T2DM rats were divided into five groups at random: (1) sham group (sham-operated rats, n = 14); (2) PBS group (T2DM rats treated with MCAO and PBS, n = 14); (3) LV-APN-EPCs group (T2DM rats treated with MCAO and LV-APN-EPCs, n = 14); (4) LV-EPCs group (T2DM rats treated with MCAO and LV-GFP-EPCs, n = 14); (5) EPCs group (T2DM rats treated with MCAO and EPCs, n = 14).

Fifty-six T2DM rats were subjected to transient (2 h) left MCAO via intraluminal vascular occlusion, as previously described (20). In brief, A nylon filament (heparin-dampened, 0.30 mm diameter, Beijing Cinontech Biotech Co., Ltd., Beijing, China) was put into the left common carotid artery (CCA) lumen and softly inserted into the internal carotid artery (ICA) for about 18 mm. Two hours later, reperfusion was initiated by thread withdrawal. Fourteen T2DM rats were subjected to sham operations; the procedure was the same as that of MCAO except for the insertion of a nylon filament. The exclusion criteria were MCAO rats on a five-point scale (“0”: no apparent deficits; “1”: left forelimb flexion; “2”: a decreased grip of the left forelimb while the tail was pulled; “3”: spontaneous movement in all directions; “4”: spontaneous left circling) (21). Rats who score <1 (possibly small to no lesion) or > 3 (poor survival) at 2 h after MCAO (before treatment) were excluded. At 1 h after reperfusion, LV-APN-EPCs, LV-EPCs, and EPCs resuspended in PBS were injected into the tail vein of rats (1 × 10⁶/ml per rat, LV-APN-EPCs group, LV-EPCs group and EPCs group), whereas 1 ml of PBS was IV-administered to the rats (PBS group and Sham group). T2-weighted imaging (T2WI) was performed to verify the ischemic stroke after treatment.

The modified neurological severity score (mNSS) was evaluated before MCAO and on days 1, 7, and 14 after MCAO by an observer, who was blinded to the groups. Table 1 shows the mNSS points (22). Neurological function was graded as 0–18 (0, normal score; 18, most severe score).

All of the animals underwent brain MRI scans using a 7.0-T small-animal MR scanner (Bruker PharmaScan, Ettlingen, Germany) on day 14 post MCAO to measure the infarct rate. Brain scanning was conducted through T2WI sequence with the following parameters: a matrix of 256 × 256, echo time (TE) = 36 ms, repetition time (TR) = 3,000 ms, field of view (FOV) = 3 × 3 cm², a total of 24 slices with a slice thickness of 0.8 mm, and the total sequence time was ~9 min. The infarct rate was determined on T2WI with the use of Image-Pro plus 6.0 software (Media Cybernetics Inc., Bethesda, MD, USA). The calculation of infarct rate was derived from the following formula (23): Infarct rate = (area of the control hemisphere—area of the non-infarcted region in the lesioned hemisphere)/area of the control hemisphere × 100%.

After brain scanning, the rats were anesthetized with 3% pentobarbital sodium. Five rats were randomly selected from each group and transcardially perfused with 4% paraformaldehyde solution and sacrificed using decapitation. The brains were immediately removed and post-fixed overnight in 4% paraformaldehyde and embedded in paraffin after full dehydration. Then, the brains were cut at 4-μm-thick coronal sections and the sections were dewaxed before use. H&E staining was done to view the pathological changes in the peri-infarct region of T2DM stroke rats. Abnormal neurons were shrunk and deeply stained, while normal neurons were orderly arranged with normal morphology and evident nucleus and nucleolus. We observed pathological changes by counting the number of viable normal neurons in the field.

Cellular apoptosis was detected using immunofluorescent staining of TUNEL assays in the peri-infarct cortex. It was performed following the manufacturer's instructions (Yeasen, Shanghai, China). TUNEL-positive cells and DAPI-stained nuclei were counted by the Image-Pro Plus 6.0 software. We calculated the percentage of the positive cells using the following formula: Apoptosis rate = TUNEL-positive cells/total cells per field × 100%.

On day 14 post MCAO, five rats were randomly selected from each group and brains of the rats were taken after decollation. The cortex tissues of peri-infarct region were isolated on the ice rapidly and transferred to a refrigerator at −80°C for storage immediately. Total protein samples were extracted from the cortex of peri-infarct region, which was preserved in liquid nitrogen (−196°C). Protein concentrations were tested by using BCA Protein Assay Kit (Pierce, USA). The primary antibodies of anti-Bcl-2 (1:1,000, CST) and anti-Bax (1:1,000, Abcam) were used. GAPDH served as a loading control. In brief, 40-μg protein samples were segregated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to polyvinylidene difluoride (PVDF) membranes (Millipore, USA), which were blocked shakily with 5% non-fat milk for 2 h and incubated with the primary antibodies of anti-Bcl-2 and anti-Bax overnight at 4°C. After being incubated for a night, the membranes were washed by Tris-buffered saline and Tween-20 (TBST) five times and incubated with horseradish peroxidase-conjugated secondary antibodies (1:50,000, Boster, Wuhan, China) at 37°C for 2 h. Finally, an enhanced chemiluminescence detection kit (Pierce, USA) was used to develop, and BandScan (Glyko, ProZyme, USA) was used to analyze the bands.

Immunofluorescent staining was carried out for CD31 (a marker of endothelial cells, Abcam, Cambridge, MA, USA) on day 14; the microvessel counts were counted in the cortex of peri-infarct region from three random slices of each rat brain and five rats from each group. We counted the number of the CD31-positive cells per field to evaluate microvessel density in the cortex of the peri-infarct region by fluorescence microscopy. Two investigators who were blinded to the groups carried out this quantification.

The levels of eNOS and vascular endothelial growth factor (VEGF) in the ischemic brain were quantified with ELISA kits, according to the manufacturer's protocols (Nanjing, China).

All statistical analyses were performed by using GraphPad Prism (version 7.0; GraphPad Software, La Jolla, CA) and SPSS 22.0 software (IBM Corp, Armonk, NY, USA). The t-test and one-way analysis of variance (ANOVA) were used to test the intergroup differences in means of two and multiple groups for normally distributed variables, respectively. The Mann–Whitney U test and Kruskal–Wallis test were used to test the intergroup difference in means of two and multiple groups for non-normally distributed variables, respectively. Data were presented as mean ± standard deviation. A two-tailed P < 0.05 was considered statistically significant.

The fasting blood glucose in the rats (n = 70) of the five groups was measured before STZ injection and 2 weeks after STZ injection. There was no significant intergroup difference both before (P = 0.56) and after STZ injection (P = 0.07). The fasting blood glucose in all rats was more than 16.7 mmol/L after STZ injection (P < 0.05 vs. before STZ injection) (Figure 1). After the establishment of the T2DM model, the rats appeared with polyuria, polydipsia, polyphagia, and weight loss as well as frequent fluctuations in high blood glucose.

The MCAO model was verified by brain scanning with T2WI sequence 24 h after transplantation. The success rate of this bolt line method was almost 100%, and the ischemic regions were nearly the same. In the process of surgery, seven rats died of vagus nerve damage and excessive bleeding. After MCAO, four rats were excluded from this study according to a five-point scale. During the following 14 days, nine rats died. Eventually, 50 rats were included in this study, with 10 rats in each group.

On day 7 after the establishment of MCAO models, a significant reduction of neurological functional deficits was found in the LV-APN-EPCs treatment group, compared with the control or EPCs or LV-EPCs treatment groups [(5.60 ± 0.52) vs. (7.70 ± 0.82), (8.20 ± 1.14), (9.30 ± 0.82), P < 0.05]. On day 14, LV-APN-EPCs treatment significantly decreased neurological functional deficits, compared with the control or EPCs or LV-EPCs treatment groups [(4.50 ± 0.53) vs. (6.40 ± 0.52), (6.50 ± 0.71), (7.80 ± 0.63), P < 0.05] (Figure 2).

The infarct areas are shown as hyperintensity on T2WI (Figure 3A). LV-APN-EPCs treatment effectively decreased the infarct rate compared with the PBS treatment, LV-EPCs treatment, and EPCs treatment groups on day 14 [(14.74 ± 1.11) vs. (25.29 ± 1.88), (20.24 ± 1.71), (19.75 ± 1.37), P < 0.05] (Figure 3B).

The brains of T2DM Stroke rats after PBS treatment exhibited an edematous morphology with loose organization, decreased cell size, cell disorder, and nuclear pyknosis in H&E staining, and this condition was significantly alleviated by treatment with LV-APN-EPCs (Figure 4A). Abnormal neurons were shrunk and deeply stained, while normal neurons were orderly arranged with normal morphology and evident nucleus and nucleolus. LV-APN-EPCs treatment significantly increased the number of viable normal neurons compared with PBS treatment, LV-EPCs treatment, and EPCs treatment groups on day 14 [(183.60 ± 15.44)/mm² vs. (76.00 ± 21.13)/mm², (121.20 ± 21.58)/mm², (120.00 ± 13.17)/mm², P < 0.05] (Figure 4B).

According to the results of TUNEL labeling on day 14 after MCAO, LV-APN-EPCs treatment significantly decreased the amount of TUNEL-positive cells, compared with control or EPCs or LV-EPCs treatment [(11.57 ± 1.82)% vs. (32.89 ± 4.13)%, (19.46 ± 2.56)%, (21.65 ± 1.97)%, P < 0.05] (Figure 5).

The expressions of antiapoptotic protein Bcl-2 and proapoptotic protein Bax were tested in the peri-infarct cortex on day 14 after stroke to further research how LV-APN-EPCs treatment suppressed cellular apoptosis. Western blotting results showed that LV-APN-EPCs treatment significantly increased the expression of Bcl-2 protein compared with the PBS group [(0.48 ± 0.03) vs. (0.06 ± 0.01), P < 0.05] and reduced the expression of Bax when compared with the PBS group [(0.18 ± 0.01) vs. (0.64 ± 0.03), P < 0.05] (Figure 6). These results indicated that LV-APN-EPCs treatment may inhibit cellular apoptosis through regulating the expression of apoptosis-related proteins in the peri-infarct cortex.

Compared to the PBS or EPCs or LV-EPCs treatment groups, a higher level of microvessel density in the peri-infarct region of the cortex was observed in the LV-APN-EPCs treatment group on day 14 after MCAO establishment [(80.20 ± 16.22), (138.40 ± 9.15), (125.00 ± 11.29) vs. (209.00 ± 9.25), P < 0.05] (Figure 7).

Compared with the PBS or EPCs or LV-EPCs treatment groups, significantly higher levels of eNOS expression based on ELISA assessment were detected in the LV-APN-EPCs treatment group on day 14 after MCAO establishment [(475.13 ± 145.69), (785.71 ± 103.07), (786.14 ± 103.84) vs. (951.25 ± 20.99) pg/ml, P < 0.05] (Figure 8A) and similarly higher levels of VEGF expression were detected in the LV-APN-EPCs treatment group [(101.15 ± 7.19), (169.00 ± 9.33), (146.65 ± 18.78) vs. (231.03 ± 6.81) pg/ml, P < 0.05] (Figure 8B).