A total of 30 male mice (weighing 25-30 g) were purchased from Shanghai SLAC Laboratory Animal Co., Ltd., and were exposed to artificial light from 6.00 a.m. to 8.30 p.m. in rooms with a controlled humidity (50%) and temperature (22°C) for 7 days before being used in the experiments. The animals were housed under pathogen-free conditions and were provided with ad libitum access to a constant pellet diet and water. The experiments were carried out on mice aged 7 and 10 weeks. Following the completion of experiments, the mice were euthanized by CO₂ asphyxiation at a flow rate of 20% chamber volume per min which was immediately followed by decapitation. All animal experiments and animal care were conducted in accordance with the protocols approved by the Institutional Animal Care and Use Committee of the Second Xiangya Hospital, Central South University, Changsha, China. All animal research was performed in accordance with the ARRIVE guidelines (28).

The terminal deoxynucleotidyl transferase (TdT) dUTP nick-end labeling (TUNEL) assay kit (cat. no. ab66110), antibodies against Wnt family member 3a (Wnt3a; cat. no. ab219412, 1/1,000), basic fibroblast growth factor (bFGF; cat. no. ab92337, 1/1,500), glycogen synthase kinase-3β (GSK-3β; cat. no. ab32391, 1/1,000), phosphorylated (p-)GSK-3β (cat. no. ab75745, 1/1,000), glyceraldehyde-3-phosphate dehydrogenase (GAPDH; cat. no. ab8245, 1/1,000) and horseradish peroxidase (HRP)-labeled goat anti rabbit IgG (cat no. ab205719, 1/2,000) were obtained from Abcam. Rabbit antibody against vascular endothelial growth factor (VEGF; cat. no. sc-7269, 1/1,000) was obtained from Santa Cruz Biotechnology, Inc.; the polyvinylidene fluoride (PVDF) membrane (cat. no. IPVH00010) was purchased from MilliporeSigma; the Novex™ ECL Chemiluminescent Substrate Reagent kit (cat. no. WP20005), TRIzol^® reagent (cat. no. 15596026), the RevertAid First Strand cDNA Synthesis kit (cat. no. K1621), TaqMan Pre-Amp Master Mix (cat. no. 4391128) and Lipofectamine™ RNAiMAX Transfection Reagent (cat. no. 13778030) were all purchased from Themo Fisher Scientific, Inc.; the ultrasensitive chemiluminescence imaging system (Chemi Doc™ XRS+) and fluorescence quantitative PCR (CFX Connect™) instruments were obtained from Bio-Rad Laboratories, Inc.

Brain tissue obtained from the mice was homogenized using an electric homogenizer and the total RNA was extracted using TRIzol^® reagent (Themo Fisher Scientific, Inc.) according to the supplier's protocols. cDNA was then reverse transcribed from 1 µg total RNA using the RevertAid First Strand cDNA Synthesis kit according to the manufacturer's instructions. The reverse transcription reactions were carried out under the following conditions: 42°C for 60 min; 25°C for 5 min; and 70°C for 5 min. RT-DHFRL1-4 was quantified by PCR reactions carried out under the following cycling conditions: Denaturing for 10 min at 95°C, followed by 18 cycles, each one consisting of 15 sec at 95°C and 4 min at 60°C, on a CFX Connect PCR system. The qPCR mixtures had a final volume of 20 µl and contained 1 µl of template cDNA, 0.5 µl of each sense (5′-TTA CCC AAA TAA AGT ATA) and antisense primer (5′-ATG GGT GTT GAG CTT GAA), 10 µl of TaqMan pre-amp master mix, and 8 µl of diethyl pyrocarbonate (DEPC)-treated water. β-actin was used as an internal control with (forward primer, 5′-GAC GTT GAC ATC CGT AAA GAC C; and reverse primer, 5′-CTA GGA GCC AGG GCA GTA ATC T. Each reaction was repeated for at least three times, independently. Relative mRNA levels were calculated using the 2^−ΔΔCt method (29).

The mice were randomly divided into the control (no treatment, n=6), sham-operated (sham; subjected to a procedure similar to DMCAO, although the carotid arteries were not ligated with a nylon suture, n=6) and the I/R model (subjected to DMCAO and carotid artery ligation, n=18) group. The I/R model group was further divided into the si-NC (I/R model infected with si-NC RNA, n=6) and the si-DHFRL1-4 (I/R model infected with si-DHFRL1-4 RNA, n=6) groups.

si-NC (5′-TUC UCU TGC TUG UCA UAC UTT-3′) and si-DHFRL1-4 (5′-CCU CCU GUC UUG UUC UAC UTT-3′) were obtained from Nanjing KeyGen Biotech Co., Ltd. to knockdown DHFRL1-4. Mice (n=6) were anesthetized with pentobarbital sodium (30 mg/kg, Roiland Dingchang Technology, Co., Ltd.) and fixed on a 69100 Rotational Digital Stereotaxic Instrument (RWD Life Science Co., Ltd.) to inject RNA. Before the injection, 50 µl lentivirus si-NC or si-DHFRL1-4 (10⁹ units/ml) were mixed with an equal volume of Lipofectamine RNAiMAX transfection reagent according to the manufacturer's instructions and injected into the lateral ventricle (0.2 mm posterior to the bregma, 1.0 mm lateral to the midline and 1.5 mm below the brain surface) using a microinjector equipped with the stereotaxic instrument at an injection rate of 10 µl/min. After 24 h later, the mice were used to construct the model of I/R.

The cerebral I/R model was constructed as previously described (30). Briefly, the mice were placed on heating pads to maintain their core body temperature at 37°C after being anesthetized with pentobarbital sodium (30 mg/kg) for the surgery. Distal MCAO (DMCAO) was used to generate focal cerebral ischemia. To perform DMCAO, a smooth incision was made along the middle line of the neck to bluntly separate the left sternocleidomastoid and cervical muscles and to expose the right common carotid artery. The blood vessels were pulled out to the trigeminal nerve to expose the carotid arteries. The right common carotid artery and the proximal end of external carotid artery were ligated. A small incision was made on the right common carotid artery at a distance of 1 cm from the trigeminal point and a silicon-coated nylon suture (cat. no. 602156PK5Re, Doccol Corporation) was inserted into the carotid artery to block the blood flow. The blood flow was monitored using a RFLSI III Laser Speckle Contrast Imaging System (RWD Life Science Co., Ltd.). After 2 h, the suture was withdrawn to restore the blood reperfusion for 24 h. Cefazolin antibiotics (25 mg/kg; MilliporeSigma) were subcutaneously administered immediately following surgery, and slow-release buprenorphine (1 mg/kg; ZooPharm) was subcutaneously administered 24 h following surgery. The mice in the sham-operated (sham) group were treated in an identical manner, except that DMCAO was not performed with a silicon-coated nylon suture.

At 24 h after reperfusion, the neurologic deficit was carefully scored by a neurologist (one of the authors, WM) who was blinded to the animal treatments according to a previously described method (31). Briefly, the mice were held up gently by the tail one meter above the floor and observed for forelimb flexion. The mice that extended both forelimbs toward the floor and had no other neurological deficit were assigned grade 0, mice with any amount of consistent forelimb flexion and no other abnormality were graded as 1. Mice with decreased resistance to lateral push without and with circling were graded as 2 and 3.

TTC staining was used to visualize the infarct area. On day 7, the mice were deeply anesthetized by an intraperitoneal injection of 200 µl of sodium pentobarbital and were transcardially perfused with ice-cold PBS with 0.1% diethylpyrocarbonate (MilliporeSigma) for 3 min. The sodium pentobarbital method was used only for extracting brain tissue for TTC staining, as this was milder procedure and the aim was to maintain the brain tissue unaltered for TTC staining. The mice were decapitated and the brains were frozen at -80°C for 60 min to dissect the cerebellum, brain stem and olfactory bulb. The brain tissues were sliced using a matrix device (Zivic Instruments) into 2-mm-thick coronal sections. The sections were incubated in 2% (TTC; MilliporeSigma) solution in the dark at 37°C for 15 min. During this period, the brain slices were turned round every 5 min. The TTC-stained brain sections were imaged using a digital camera (Sony HDR-PJ790, Sony Corporation) for infarct size analyses. The infarct areas were traced using ImageJ software (version 1.2.1, National Institutes of Health) as previously described (32).

H&E staining was used to visualize tissue damage as previously described (33). Briefly, the brain tissues were dehydrated through an ethanol gradient from 70 to 100% and cleared with three changes of xylene (1 min for each change). The dehydrated tissues were embedded in paraffin, sectioned to a thickness of 20 µm, dewaxed in xylene overnight and subsequently passed through 3×10 min 99% ethanol and 2×10 min 96% ethanol and rehydrated. The sections were immersed in an aqueous hematoxylin solution (cat. no. H9627, MilliporeSigma) for 3 min at 25°C, differentiated with hydrochloric acid for 15 sec, rinsed briefly and counterstained with eosin (cat. no. 318906, MilliporeSigma) for 3 min at 25°C. After rinsing thoroughly in distilled water and dehydrated as described above, the slides were cleared with xylene, sealed and examined under a microscope (Olympus CX33, Olympus Corporation).

TUNEL assay was carried to examine apoptosis in the brain tissue as previously described (34). Briefly, paraffin-embedded tissue sections were baked at 65°C for 2 h and rehydrated by passing an ethanol serial gradient and incubated in 1 µg/ml proteinase K (cat. no. 25530049, Thermo Fisher Scientific, Inc.) in a 10 mM Tris solution (7.5 µl of 20 µg/µl proteinase K in 150 ml 10 mM Tris, pH 7.4-8.0) at room temperature for 15 min. The slides were then rinsed with 1X PBS three times and blot-dried. Subsequently, 100 µl TUNEL reaction mixture were applied to each slide and the slides were incubated at 37°C in the dark for 1 h in a humid chamber according to the supplier's protocols. The slides were then washed three times with PBS, immersed in 100 mM Tris buffer (pH 8.2) at room temperature for 5 min, followed by the addition of 100 µl substrate solution and incubation in the dark at room temperature for 10 min. The nuclei were counterstained with 2-(4-amidinophenyl)-6-indolecarbamidine dihydrochloride (DAPI; cat. no. MBD0015, MilliporeSigma) at 25°C for 1 h. The slides were with then sealed with quenching solution and observed under a fluorescence microscope (Olympus BX51, Olympus Corporation).

For 10 mg tissue, 300 µl ice-cold radioimmunoprecipitation assay (RIPA) buffer (MilliporeSigma) were added and the tissue was homogenized using an electric homogenizer and agitated for 2 h at 4°C. The lysates were centrifuged at 16,000 × g for 20 min at 4°C and the proteins in the supernatant were quantified using the BCA protein assay kit (Thermo Fisher Scientific, Inc.) according to manufacturer's instructions. After denaturing by boiling at 100°C for 5 min, ~50 µg proteins were loaded onto a 12% sodium dodecyl sulfate polyacrylamide gel and separated by electrophoresis at a constant voltage of 50 V. The proteins were then transferred to PVDF membranes by electroblotting in an ice bath at a constant voltage of 24 V. The PVDF membranes were then blocked with 5% non-fat milk in 1X Tris-buffered saline with 0.1% Tween-20 (TBST) for 4 h at room temperature and incubated overnight with primary antibodies against VEGF, Wnt3a, bFGF, GSK3β, p-GSK3β and GAPDH at 4°C. The membranes were washed six times with Tris-buffered saline buffer, then incubated in HRP-conjugated goat anti-rabbit IgG (H+L) solution at room temperature for 2 h. The membranes were washed six times with PBS at room temperature and immunoreactive bands were visualized using the Novex ECL chemiluminescent substrate reagent kit in the dark according to the supplier's protocols. To quantify the proteins, the gray values of the bands were analyzed using Quantity One (version 4.62) analysis software (General Electric).

All data are expressed as the mean ± standard error of the mean (SEM) and were obtained from at least three independent assays. One-way analysis of variance with Tukey's post hoc test was used to investigate the differences among the various groups. Non-parametric ordinal data obtained for the neurologic deficit scores were analyzed using the Kruskal-Wallis test followed by Dunn's post hoc tests. A value of P<0.05 was considered to indicate a statistically significant difference.

The present study first assessed the expression of DHFRL1-4 in brain tissue after I/R modeling. The RT-qPCR data revealed that the mRNA level of DHFRL1-4 was significantly upregulated following in the I/R model group as compared with the control and sham groups (Fig. 1). However, following the injection of si-DHFRL1-4, the mRNA level of DHFRL1-4 was significantly decreased, although it was still slightly higher (although not significantly) than that of the control and sham groups. On the other hand, si-NC and sham operation did not markedly affect the DHFRL1-4 mRNA level (Fig. 1).

TTC staining revealed that following I/R modeling using the DMCAO procedure, the infarct area was significantly greater in the I/R group as compared to the control and sham groups, which di not exhibit an unstained area (Fig. 2). The injection of si-DHFRL1-4 before the I/R modelling significantly reduced the infarcted area as compared to the model group without the injection (P<0.05). On other hand, the injection of si-NC did not reduce the infarct area as compared to the model group. Similarly, I/R modelling significantly deteriorated nerve function, leading to higher neurological deficit scores in the I/R model group as compared to the control and sham groups (Fig. 2B); however, the knockdown of DHFRL1-4 significantly improved nerve function (Fig. 2B). Of note, no improvement was observed when si-NC was used (P<0.05; Fig. 2B).

The present study then assessed brain tissue injury using H&E staining. The results revealed that in the control and sham groups, the cells were arranged in a regular manner, with a relatively uniform size and intact membranes. The nuclei and nucleoli were clear and visible (Fig. 3). In the model group however, some necrosis was observed in the nerve fibers, and the fibers were swollen and had edema. The number of nerve cells was less in the I/R model group, than that in the control and sham groups with large cell-cell gaps. However, following DHFRL1-4 knockdown, there was markedly less swelling and edema as compared to the model group; the brain tissue also exhibited less liquefaction and degeneration, and the distribution of nerve cells was relatively uniform (Fig. 3).

To further investigate the mechanisms underlying the reduction of I/R injury by the knockdown of DHFRL1-4, apoptosis in the brain tissue was assessed using TUNEL assay. Compared with the control and sham groups, there were significantly more apoptotic cells after I/R modelling (P<0.05). However, apoptosis was significantly reduced when the mice were injected with si-DHFRL1-4 before I/R modelling compared with the model group, although this was still higher than that in the control and sham groups (P<0.05). On the other hand, si-NC did not reduce apoptosis (P>0.05; Fig. 4).

Since bFGF and VEGF are key growth factors related to angiogenesis, the expression of these two genes was examined in the brain tissue. Western blot analysis revealed that the expression levels of bFGF and VEGF were significantly increased after I/R modelling (P<0.05). DHFRL1-4 knockdown further upregulated these expression levels (P<0.05); however, si-NC did not alter bFGF and VEGF expression after I/R modelling (P>0.05; Fig. 5).

In addition, western blot analysis revealed that the protein levels of Wnt3a, GSK-3β and p-GSK-3β were significantly upregulated after I/R modelling as compared with the control and sham groups (P<0.05 and P<0.01). After DHFRL1-4 knockdown, the Wnt3a and GSK-3β levels were reduced; however, the p-GSK-3β level was not altered, resulting in an increased p-GSK-3β/GSK-3β ratio (P<0.01; Fig. 6). On the other hand, si-NC did not affect the protein levels compared with the I/R model group (P>0.05; Fig. 6).