The animal experiments were carried out after being approved by the ethics committee of Laboratory Animal Center of Nantong University (NTU) (Ethics Approval No. IACUC20221230-1006) and according to Animal Care and Use Guidelines, and every attempt was made to minimize discomfort and pain to the animals.

The CIRI model was constructed as depicted in a prior study [22] with slight modifications. 8-week-old C57BL/6 mice (male, 22–26 g, n = 160) were ordered from Hangzhou Medical College (Zhejiang, China) and maintained in the specific pathogen-free cages at 26 °C and 50–70% relative humidity. All mice could get standard experimental food and sterile tap water at will and were raised following the circadian rhythm.

The process of restoring blood flow and causing reperfusion injury after cerebral ischemia was called CIRI modeling. For the establishment of animal CIRI model, after being fasted for 12 h, all mice were inhaled with 2% isoflurane (I40690, Acmec Biochemical, Shanghai, China) and anesthetized, with fur and skin pre-disinfected using povidone-iodine (A606166, Sangon Biotech, Shanghai, China) in a supine position. The body temperature was retained at 37.0 $\pm$ 0.5 °C with a heating lamp. The neck was incised along the midline to expose the internal carotid artery (ICA), external carotid artery (ECA) and common carotid artery (CCA). The ligation and electro-coagulation of ECA were executed, while the CCA and ICA were clamped at the same time, subsequent to which an oblique incision within the ECA was created. The 5 mm end of the No. 3 monofilaments nylon fishing line was wrapped in paraffin (P46564, Acmec Biochemical), and a mark was made at the length of 18 mm. Then the fishing line was inserted into incision at approximately 18–20 mm from the bifurcation of the CCA, and the middle cerebral artery (MCA) of right brain was embolized. The skin was sutured while the end of the fishing line was partially fixed to the skin. After the induction of ischemia for 30 min, the fishing line was carefully withdrawn around 8 mm so as to construct the reperfusion model. The mice in the Sham group received the insertion of nylon fishing line around 10 mm without the embolization, and the remaining steps were the same as those in the modeling group. After the animals were awake after surgery and their vital signs were stable, that is, 1 h after the modeling, they were graded using neurological deficit score by researchers who were blinded to the grouping and obtained 1–3 scores, the modeling was considered successful [23]. Then they were raised in individual cages at 26 °C and had free access to food and water.

The specified criteria were listed as follows: grade 0: there was no defect in mice; grade 1: mice were incapable of straightening the contra forelimb contralateral to the embolized artery when their tails were lifted; grade 2: mice circled to the contralateral side of the embolized artery while walking; grade 3: mice titled to the opposite side of the embolized artery while walking; grade 4: there was severe defect without spontaneous motor activity or consciousness.

To determine the change on the endogenous expression level of VDR in mice, mice were stochastically distributed to the groups of Sham, ischemia (30 min), and reperfusion (3, 6, 12 and 24 h) (n = 10/group). For intervention assay to unravel the effect of activating VDR in CIRI-modeled mice, mice were assigned to the groups of Sham, CIRI, Vehicle, Calcitriol and Paricalcitol (PC) (n = 10/group), and those in Vehicle, Calcitriol and PC groups were given the solvent including 95% propylene glycol (P62861, Acmec Biochemical) and 5% ethanol (E99942, Acmec Biochemical), Calcitriol (a natural agonist of VDR, 1 µg/kg, C39241, Acmec Biochemical) and PC (a synthetic agonist of VDR, 1 µg/kg, P44940, Acmec Biochemical) via intraperitoneal injection 15 min before the reperfusion [17]. Subsequently, two schemes were applied for the analyses: mice were either subjected to ischemia for 30 min and reperfusion for 3 h to determine the apoptosis of cerebral cells or subjected to the ischemia for 30 min and reperfusion for 24 h to detect the infarct volume.

Additionally, the mice were distributed to the groups of Sham, CIRI, Vector, short hairpin RNA against VDR (shVDR) and VDR (n = 10/group). The mice in the last three groups received intraventricular injection of adenoviruses of vector (control), shVDR and VDR overexpression plasmid 2 h prior to the construction of CIRI model [24]. Also, to determine the apoptosis of cerebral cells and the infarct volume, mice were subjected to the scheme as mentioned above.

At the end of the experiment, all mice were anesthetized by 1% sodium pentobarbital (45 mg/kg, P3761, Sigma-Aldrich, St. Louis, MO, USA) and euthanized with neck dislocation. Then the brain tissue was removed and snap-frozen in liquid nitrogen until usage.

All procedures here referred to previous studies [17, 24]. The adenoviral shuttle vector pDC316-mCMV-EGFP (FH1691, FengHui Biotechnology, Co., Ltd., Hunan, China) and backbone plasmid pBHGloxdeltaE13Cre (JiRan Biotechnology, Shanghai, China) were prepared in advance. The cDNA of shVDR and VDR was cloned into the shuttle vector to obtain pDC316-mCMV-shVDR-EGFP and pDC316-mCMV-VDR-EGFP. The adenoviruses of VDR and shVDR were created in accordance with the protocol of an adenoviral system (632269, Takara Bio, Osaka, Japan), and the viral titers were assayed by Adeno-X-Rapid Titer kit (632250, Takara Bio). The virus only carrying empty coding sequences was used as the control.

For adenovirus-mediated overexpression or knockdown of VDR, the adenoviruses were dissolved in 0.9% saline (ST341, Beyotime, Shanghai, China) and administrated into mice 2 h prior to CIRI modeling. Specifically, after the mice were anesthetized and placed on 71000-M automatic stereotactic head frame (RWD Lifescience, Guangdong, China), the adenoviruses were slowly infused (0.2 µL/min) into the right lateral ventricle with the help of an microinfusion pump (VOGT Medical, Karlsruhe, Germany). Bone wax was used to seal the burr hole for the prevention of possible leakage. The suture of incision was made and mice were put back to cages for recovery, with free access to water and food.

To determine the infarct volume, the collected tissue was rinsed with saline for 5 min and placed at –20 °C for 20 min, followed by being transferred to a brain slice mold. The remaining tissues were perfused with 0.9% saline, fixed with 4% paraformaldehyde (P39200, Acmec Biochemical), embedded in paraffin and sliced as required [25].

The infarct volume in the brains of mice was determined as follows [26]. The mouse brains in each group were coronally sectioned into 2 mm-thick slices, dyed with 2% 2,3,5-triphenyl-tetrazolium chloride (TTC, T54930, Acmec Biochemical) at 37 °C for 20 min, and finally fixed with 10% formalin (AC11741, Acmec Biochemical) for 24 h. A digital camera (D500, Nikon, Tokyo, Japan) was used to photograph the brain sections and the infarcted area was quantified via ImageJ (v. 1.8, Bio-Rad, Hercules, CA, USA). The infarct volume of mice was determined by the formula:

$$infarctvolume(\%)=\frac{infractedarea}{totalbrainarea}\times 100\%$$

The apoptosis of cerebral cells was detected as depicted in a previous study [27]. The TUNEL staining was performed following the producer’s protocols. In detail, the 5 µm-thick sections of fixed paraffin-embedded brain tissue were obtained, following which the sections were deparaffinized and rehydrated. The sections were then immersed in 3% hydrogen peroxide (H112515, Aladdin, Shanghai, China) and in 0.1 M sodium citrate (S25420, Acmec Biochemical) in a microwave for 5 min, followed by treatment with TUNEL assay kit (11684795910, Roche Diagnostics, Indianapolis, IN, USA) and dyeing with diaminobenzidine (DAB, AC11043, Acmec Biochemical) in sequence. The TUNEL-positive cerebral cells were observed in a fluorescence microscope ($\times$200, BX53, Olympus, Tokyo, Japan).

Human neuroblastoma cell line SK-N-SH (AW-CELLS-H0335, AnWei Biotechnology, Shanghai, China) was maintained in minimum essential medium (iCell-0012, iCell, Shanghai, China) blended with 10% fetal bovine serum (iCell-0500, iCell), 1% non-essential amino acid (iCell-01000, iCell), 1% L-GlutaMax (iCell-0900, iCell) and 1% penicillin-streptomycin (iCell-15140-122, iCell). CB-S Solid.Line 170 incubator (BINDER, GmbH, Tuttlingen, Germany) was used to culture the cells with 5% CO₂ at 37 °C as needed. Short tandem repeat method was exploited to verify the cell line. The cells used in this study were validated by STR profiling and tested negative for mycoplasma.

To mimic CIRI in vitro, SK-N-SH cells were treated with the OGD/R as described [28]. Briefly, the normal medium for cell culture was replaced by glucose-free Dulbecco’s modified Eagle’s medium (iCell-138-0001, iCell) and cells were incubated in an incubator at 37 °C with 95% N₂ and 5% CO₂ for 4 h. For reoxygenation process, cells were then maintained in normoxic condition (95% air and 5% CO₂) for 24 h. Cells in the control group were cultured normally without OGD/R.

To determine the effect of VDR in SK-N-SH cells, the cells were transfected with the shVDR (C02002) and its control (Vector, C03002) obtainable from GenePharma (Shanghai, China) with the help of lipofectamine RNAiMAX transfection reagent (13778150, Thermo Fisher Scientific, Waltham, MA, USA). The sequences used were available in Table 1.

Then cells were treated with OGD/R only or additionally treated with N-acetyl-L-cysteine (NAC, N27390, Acmec Biochemical), a scavenger of oxidative stress, at 5 mM for 3 h [29].

2 $\times$ 10³ cells/well SK-N-SH cells were seeded into a 96-well plate and incubated for 24 h. 10 µL CCK-8 solution (CA1210, Acmec Biochemical) was added for 4-h additional incubation. The iMark microplate reader (E1140, Beyotime) was used to read the value of optical density (OD) at an absorbance of 450 nm. Cell viability (%) was calculated with the formula:

$$Cell viability(\%)={\displaystyle \frac{[OD(dosage)-OD(blank)]}{[OD(nodosage)-OD(blank)]}}\times 100\%$$

The apoptosis of SK-N-SH cells with indicated treatment was determined via flow cytometry using a commercial kit (C0162S, Beyotime). After the centrifugation at 1000 $\times$g for 5 min, the supernatant was removed and cells were subsequently resuspended in phosphate-buffered saline (PBS, C0221A, Beyotime). 1 $\times$ 10⁶ resuspended cells were centrifuged for 5 min at 1000 $\times$g, and the supernatant was removed, following which 195 µL Annexin Ⅴ-FITC binding buffer was added. 5 µL Annexin Ⅴ-FITC and 10 µL propidium iodide were introduced to the cells and mixed for 20-min incubation at room temperature without light. CyFlow Cube 6 (Sysmex Corporation, Kobe, Japan) and Kaluza C (v. 2.1, Beckman-Coulter, Indianapolis, IN, USA) were employed to analyze the results.

A ROS assay kit (S0033S, Beyotime) was utilized to estimate the level of ROS in the brain tissue slices (5 µm-thick) and SK-N-SH cells (3 $\times$ 10³ cells/well in 96-well plates) were treated or transfected as described [30]. In detail, the tissue slices and the cells were both immersed in the dichlorodihydrofluorescein diacetate (DCFH-DA) at the final concentration of 10 µM for 30 min at 37 °C, then rinsed and mounted by the fluorescent mounting media (P0126, Beyotime). A fluorescence microscope (Eclipse TE-300, Nikon) was used to photograph the results under the magnification of $\times$100, and the level of ROS was quantified by ImagePro Plus 6.0 (Media Cy, Bethesda, MD, USA).

Two schemes were adopted for the analyses on the levels of factors related to oxidative stress [30, 31]. In the first scheme, the tissues harvested from the cerebral peri-infarcted areas of mice were homogenized mechanically and centrifuged for 10 min to collect the supernatant. Following the detection of protein concentration by bicinchoninic acid protein assay kit (AC13858, Acmec Biochemical), the content of malondialdehyde (MDA, KTE71091, Abbkine, Hubei, China) and the level of superoxide dismutase (SOD, CSB-E08556m, Cusabio, Fannin, TX, USA) were determined by a microplate reader according to the protocols.

In the second scheme, SK-N-SH cells (1 $\times$ 10⁵ cells/mL) of each group were incubated in a 90-mm culture dish for 24 h, rinsed with ice-cold PBS, centrifuged at 1000 $\times$g for 5 min, transferred into 0.5 mL precooled PBS and homogenized. The homogenate was additionally centrifuged for 15 min at 8000 $\times$g, and the concentration of protein was measured via bicinchoninic acid protein assay kit. The supernatant was collected for the assays of MDA (XPEH4174), SOD (XPEH4706-1) and catalase (CAT, XPEH0643) by a microplate reader (XpressBio, Frederick, MD, USA).

The total RNA in the tissue of cerebral peri-infarcted areas and SK-N-SH cells were harvested by Trizol (R0016, Beyotime) and then CM-36dG spectrophotometer (Konica Minolta, Tokyo, Japan) was used to assess the quality and quantity of the RNA. cDNA was synthesized via a QuantiTech Reverse Transcription kit (205311, Qiagen, Hilden, Germany), and the PCR was operated by a multiplex PCR kit (206145, Qiagen) in a PCR system (911001, Qiagen) under the following conditions as per the protocols:

(1) initial heat activation: 95 °C for 15 min

(2) 3-step cycling in a total number of 40 times:

a. 94 °C for 30 seconds (s)

b. 60 °C for 90 s

c. 72 °C for 90 s

(3) 72 °C for 10 min.

Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was the internal control and 2^{-Δ⁢Δ⁢CT} method was employed to analyze the relative expressions [32].

Primers were listed below:

(mouse) VDR: Forward: 5^′-ATTCGTGCAGACGTAAGTA-3^′

Reverse: 5^′-GGTGTACAGATCAGAGTTTG-3^′

(human) VDR: Forward: 5^′-ATGAAGGAGTTCATTCTGAC-3^′

Reverse: 5^′-CTGAATCCTGGTATCATCTT-3^′

(mouse) GAPDH: Forward: 5^′-CCTCAACTACATGGTTTACA-3^′

Reverse: 5^′-TGTTGTCATACTTCTCATGG-3^′

(human) GAPDH: Forward: 5^′-GATGCTGGTGCTGAGTAT-3^′

Reverse: 5^′-CACAAAGTTGTCATTGAGAG-3^′

All reagents, except for those indicated, were the products of E-Labscience (Hubei, China). Radio immunoprecipitation assay lysis buffer which contained phosphatase inhibitors and protease (E-BC-R327) was utilized for the extraction of proteins in cerebral tissue and SK-N-SH cells. Following the separation on 12% gel reagent (E-IR-R305), the samples were transferred to the polyvinylidene fluoride membranes (E-BC-R266). Subsequently, the membranes were soaked in 5% defatted milk for 60 min and, under the condition of 4 °C overnight, were incubated with the primary antibodies (1:1000; Cell Signaling Technology, Danvers, MA, USA) against VDR (#12550, 54 and 48 kDa), Bcl-2-associated X (Bax, #2772, 20 kDa), B-cell lymphoma-2 (Bcl-2, #3498, 26 kDa), Cleaved Caspase-3 (#9661, 19 and 17 kDa), and housekeeping gene GAPDH (#5174, 37 kDa). After that, the membranes were rinsed using TBST (E-BC-R347) and continuously incubated with horseradish peroxidase conjugated goat anti-rabbit IgG (E-AB-1003, 1:5000) at room temperature for 60 min. Enhanced chemiluminescence reagent (E-BC-R347) was applied to detect the protein bands, and the chemiDoc imaging system (Bio-Rad) with Quantity One software 4.6 (Bio-Rad) was employed to analyzed the results [33].

GraphPad Prism 8 (GraphPad, Inc., San Diego, CA, USA) was utilized to analyze the data of three independent tests to reduce errors and obtain the mean value. All data were displayed as mean $\pm$ standard deviation (SD). Student’s t test and one-way ANOVA were conducted for two-group and multi-group comparisons, the normality test results for all data are shown in the Supplementary Table 1. The Bonferroni post hoc test was adopted for multi-group contrast. All the data followed a normal distribution. The Shapiro–Wilktest was applied for normality, and the Brown–Forsythe testwas applied for homogeneity of variances. A p-value below 0.05 was considered as statistical significance.

Former findings made us wonder if VDR could impact mice with CIRI, and accordingly, we measured its mRNA and protein expressions in cerebral peri-infarcted areas of mice by qRT-PCR and Western blot, and found similar results that the level of VDR was evidently increased in the cerebral tissue of CIRI mice compared to mice with 30 min ischemia (Fig. 1A,B, p 0.001). Time-course research on the expression level of VDR revealed that the level of VDR in cerebral tissue did not change significantly during ischemia, but elevated in ischemic area of cerebral tissue in CIRI mice after the reperfusion at 12 and 24 h relative to mice with 30 min ischemia (Fig. 1A,B, p 0.001), the original Western Blot (WB) images were available in the Supplementary file.

Subsequently, in order to determine whether the activation of VDR could protect against CIRI-induced cerebral injury in mice, we treated CIRI mice with vehicle (the solvent only), Calcitriol or PC, and then calculated the cerebral infarct volume based on TTC staining. A sharp increased infarct volume was seen in CIRI mice compared with Sham mice (Fig. 2A,B, p 0.001), whereas in comparison with Vehicle group, the CIRI-induced cerebral infarct volume was apparently reduced in Calcitriol and PC groups (Fig. 2A,B, p 0.001). Based on TUNEL assay, the number of TUNEL-positive cells was elevated in CIRI mice compared with Sham mice (Fig. 2C, p 0.001), whereas such elevation was neutralized via the activation of VDR in Calcitriol and PC groups compared to Vehicle group (Fig. 2C, p 0.001).

To unravel the effect of VDR in CIRI mice, we subsequently constructed the adenoviruses of both VDR overexpression plasmid and shVDR to manipulate VDR expression. In light of Western blot and qRT-PCR data, both protein and mRNA expressions of VDR were increased in CIRI mice relative to Sham mice (Fig. 3A,B, p 0.001), which was further enhanced by VDR overexpression plasmid (Fig. 3A,B, p 0.001), yet suppressed by shVDR (Fig. 3A,B, p 0.001), , the original WB images were available in the Supplementary file. Meanwhile, according to the results of TTC staining and TUNEL assay, increased infarct volume and TUNEL-positive cells were found in CIRI mice compared with Sham mice (Fig. 3C–E, p 0.001), which were aggravated following VDR silencing (Fig. 3C–E, p 0.05), but attenuated by overexpressed VDR (Fig. 3C–E, p 0.05).

Additionally, as for the levels of oxidative stress-related factors, in CIRI mice, an upward trend was determined in the levels of ROS and MDA, whilst a downward trend in the level of SOD was found in comparison to Sham mice (Fig. 4A–C, p 0.01). VDR silencing in CIRI mice further raised levels of MDA and ROS yet dropped level of SOD (Fig. 4A–C, p 0.05), whereas VDR overexpression brought about contrary results (Fig. 4A–C, p 0.05).

To further confirm the interplay between VDR and oxidative stress, we employed SK-N-SH cells to construct an OGD/R-treated cell model, and OGD/R-treated cells were transfected with shVDR together with/without treatment of NAC, a scavenger of oxidative stress. It could be seen that the level of VDR was sharply increased in OGD/R-treated SK-N-SH cells in comparison to Control SK-N-SH cells (Fig. 5A, p 0.001), which was then neutralized due to silencing of VDR (Fig. 5A, p 0.001). Additionally, we confirmed that NAC didn’t affect the VDR expression level in OGD/R-treated SK-N-SH cells with VDR silencing (Fig. 5A).

Also, in flow cytometry and CCK-8 assays, the cell viability was suppressed yet the apoptosis was enhanced in OGD/R-treated SK-N-SH cells compared with Control SK-N-SH cells (Fig. 5B–D, p 0.001), which was further strengthened by VDR silencing (Fig. 5B–D, p 0.05). Then, NAC attenuated the role of VDR silencing by promoting the viability yet hindering the apoptosis of SK-N-SH cells (Fig. 5B–D, p 0.01).

Furthermore, the results indicated that the levels of ROS and MDA were upregulated yet those of SOD and CAT were downregulated in OGD/R-treated SK-N-SH cells (Fig. 5E–H, p 0.01), and such trends were further promoted via silencing of VDR (Fig. 5E–H, p 0.01). Also, NAC was capable of reversing the effects of VDR silencing in OGD/R-treated SK-N-SH cells (Fig. 5E–H, p 0.001). Further, the elevated levels of both Cleaved Caspase-3 and Bax yet decreased level of Bcl-2 were found in OGD/R-treated SK-N-SH cells (Fig. 6A–D, p 0.01), which was further strengthened in response to the silencing of VDR (Fig. 6A–D, p 0.01). NAC, on the other hand, counteracted such effects of VDR silencing in OGD/R-treated SK-N-SH cells (Fig. 6A–D, p 0.001), the original WB images were available in the Supplementary file.