Tilianin with 99.02% purity, shown by high-performance liquid chromatography (HPLC) was isolated from D. moldavica (30, 36). Briefly, the above-ground parts of D. moldavica were ground to powder and refluxed three times with 40% ethanol at 100 °C. The combined ethanol solution was filtered and evaporated at low pressure to obtain the crude extract, which was then separated by column chromatography using HPD600 resin and eluted with water, 50%, and 70% ethanol. Impurities were removed using a silica gel column and the pure compound was collected. The structure and HPLC image of tilianin were shown in Figures 1A, B .

Male Sprague–Dawley rats aged seven weeks and weighing 220 ± 20 g were purchased from Vital River Laboratory Animal Technology Co., Ltd. (Beijing, China). A total of 136 rats were included in this study. All experiments were approved by the Experimental Animal Care and Use Committee of the Institute of Medicinal Biotechnology (No. IMB-202102-D7).

The two-vessel occlusion (2VO) model is an established model of VaD. Following the standard procedure, rats were anesthetized with pentobarbital sodium (30 mg/kg). We then separated the bilateral common carotid arteries and ligated them with a No.4–0 surgical suture. Rats in the sham group underwent the same surgery but without ligating the bilateral common carotid arteries. During the operation, the anal temperature of the rats was maintained at 37 °C. The rats were tested 21 days post-surgery.

Next, the rats were randomly divided into three first-level groups: tilianin treatment group, target overexpression or downregulation group, and target overexpression or downregulation with tilianin treatment group. Within the tilianin treatment group, rats were subgrouped into sham (n = 8), sham+40 mg/kg tilianin (n = 8), 2VO (n = 8), and 2VO+40 mg/kg tilianin (n = 8) groups. In the target overexpression or downregulation group, rats were subgrouped as sham+negative control mimics (NCM) (n = 8), sham+miR-152-3p mimics (152M) (n = 8), sham+miR-193b-3p mimics (193M) (n = 8), 2VO+NCM (n = 8), 2VO+152M (n = 8), 2VO+193M (n = 8), 2VO+152M+CaMKIIα (n = 8), and 2VO+193M+CaM (n = 8) groups. In the target overexpression or downregulation with tilianin treatment group, rats were subgrouped into sham+negative control inhibitor (NCI) (n = 8), 2VO+NCI (n = 8), 2VO+NCI+tilianin (n = 8), 2VO+miR-152-3p inhibitor (152I)+tilianin (n = 8), and 2VO+miR-193b-3p inhibitor (193I)+ tilianin (n = 8) groups. Rats in each group received the corresponding operations and drugs. The experimental design is shown in Figure 1C .

Group I (negative control group) received a daily corn oil dose (5 mL/kg/body weight, BW) through a gastric tube for one month.

Rats in the tilianin treatment group received a daily dose of 40 mg/kg tilianin that was dissolved in 0.5% carboxymethylcellulose sodium (CMC-Na) (5 mL/kg/body weight) with a gastric tube for four weeks. Rats in the sham and 2VO groups were given 0.5% CMC-Na vehicle (5 mL/kg/body weight) over the same period. For the stereotactic injection procedures, rats were anesthetized with isoflurane, placed in an animal stereotaxic apparatus (RWB Life Science Co, Ltd., China), and the bregma was marked. The injection coordinates were as follows: anteroposterior, -1.0 mm; mediolateral, ± 1.5 mm; and dorsoventral, -4.5 mm. A total of 8 μL adeno-associated viruses (AAVs) were injected into the lateral ventricle. After the injection, the needle was left in place for 10 min to allow the liquid to be fully absorbed. The miR-193b-3p mimics (AAV-CAG-miR-193b-3p), miR-152-3p mimics (AAV-CAG-miR-152-3p), miR-193b-3p inhibitor (AAV-CAG-miR-193b-3p inhibitor), miR-152-3p inhibitor (AAV-CAG-miR-152-3p inhibitor), CAML (AAV-CAG-CALM), and CAMK2A (AAV-CAG-CAMK2A) were synthesized by Sangon Biotech (Shanghai, China); titer, ≥ 1×VG¹² vg/mL. The transfection efficiency of AAV-miRNAs is shown in Figure S1 .

Human neuroblastoma SH-SY5Y cells and human embryonic kidney cells (HEK293) (ATCC, USA) were cultured in high-glucose Dulbecco’s Modified Eagle Medium (DMEM) with 10% fetal bovine serum (Gibco, USA) and penicillin at 37°C in an atmosphere of 5% CO₂. The OGD injury model was used to simulate the ischemic injury associated with VaD in vitro (9). SH-SY5Y cells were subjected to OGD, similarly as previously described (37). SH-SY5Y and HEK293 cells were transfected using Lipofectamine 2000 according to the manufacturer’s protocols (Invitrogen, Carlsbad, CA, USA). SH-SY5Y cells in individual wells were transfected with 152M, 193M, NCM, NCI, pCMV6-CaMKIIα, and pCMV6-CaM (Sangon Biotech, Shanghai, China). HEK293 cells were co-transfected with recombinant PGL3 plasmids and miRNA sequences to validate the miRNA targets.

The MWM is a classic experiment used to evaluate spatial cognition and memory in rats. The test consisted of positioning navigation for the first 5 days and a probe trial on day 6. For positioning navigation, the rats were placed in the water in the four quadrants of the maze for four tests. The time taken for the rats to successfully find the hidden platform (escape latency) and swimming speed were recorded. The average time and average swimming speed of the four tests were used for the analysis. If the rats did not find the platform within 60 s, they were directed to the platform and allowed to stay there for 10 s to learn its location. In the probe trial, the rats entered the water from the quadrant furthest from the original platform. The time spent within the target quadrant (position of the original platform) and number of times crossing the original platform were recorded.

Rats were sacrificed using pentobarbital sodium and perfused with 4% paraformaldehyde. The brains were dehydrated with graded concentrations of ethanol (50%, 70%, 85%, 95%, 100%, and 100%), and then transparentized with xylene and ethanol (50% xylene + 50% ethanol, 100% xylene, 100% xylene). Next, the brains were placed in paraffin-xylene solutions (50% paraffin, 100% paraffin) at 60 °C, followed by embedding in paraffin. The embedding blocks were fixed on the specimen table, cut into 4 μm sections, and subsequently placed in warm water at 45 °C. After the sections were unfolded, they were attached to the slides. Subsequently, the sections of the cortex and hippocampus were processed for Nissl staining (Servicebio, Wuhan, China) and TUNEL staining (Norvan, Nanjing, China) according to the manufacturers’ instructions.

After blocking with 5% bovine serum albumin (BSA) for 1 h, the cortex and hippocampus sections were immunohistochemically stained for GFAP and IBA1 with primary antibodies ( Table S1 ) overnight at 4°C, followed by goat anti-rabbit IgG (H+L)-HRP secondary antibody ( Table S1 ) for 1 h at room temperature. Both primary and secondary antibodies were diluted in phosphate buffer saline containing 0.3% Triton X-100 (PBST) and 1% BSA. Histological images were collected under a bright microscope (Olympus CX33-LV2000, Tokyo, Japan) using a 20× objective with the same condition at a resolution of 1280 × 960 pixels. The percentage area of the GFAP or IBA1 positive cells in the hippocampus and cortex in each sections were calculated by ImageJ software (NIH, MD, USA). The mean percentage area were calculated from 3 random microscopic fields from each section for each animal by one person unaware of treatment history. A total of 3 animals per group were analyzed.

Brain tissues were screened for abnormal miRNAs and mRNAs by high-throughput sequencing analysis (Sangon Biotech, Shanghai, China). Differentially expressed miRNAs and mRNAs in the brain of rats in the sham, 2VO, and tilianin groups were screened using the criteria of P<0.05 and |log2Fold Change|>1. The miRDB (http://www.mirdb.org/), miRWalk (http://mirwalk.umm.uni-heidelberg.de/), and TargetScan (https://www.targetscan.org/vert_80/) databases were used to predict the target genes of the miRNAs. Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment were used to explore the molecular mechanisms and signaling pathways of VaD pathogenesis. A protein-protein interaction (PPI) network was constructed using STRING 11.0 software to visualize the PPI network of differentially expressed genes.

Total RNA in the hippocampus and cortex of rats, SH-SY5Y cells, HEK293 cells, or plasma was extracted with TRIzol reagent from Invitrogen. For mRNA, cDNA was synthesized by HiscriptIII RT SuperMix and the qRT-PCR was completed using SYBR Master Mix kits (Vazyme, Nanjing, China). For miRNA and U6, miRNA Synthesis Kit and miRNA SYBR Master Mix were used (Vazyme, Nanjing, China). Sangon Biotech (Shanghai, China) produced all primers ( Table S2 ). Each experiment contained three technical replicates and four biological replicates.

SH-SY5Y cells were cultured in 96-well plates at a density of 1×10⁴ cells/well and transfected as described above. After 24 h, the cells were cultured with 100 μL Cell Counting Lite (CCL) solution (CellCounting-Lite 2.0 Luminescent Cell Viability Assay, Novazan Co.LTD, Nanjing) for 15 min. A Spark 20M multimode micrometer (Tecan Group Ltd., Mannedorf, Switzerland) was used to determine the optical density (OD). Each assay contained three technical replicates and three biological replicates.

The degree of apoptosis in SH-SY5Y cells was determined using the Annexin V (Annexin V-PE)/7-amino-actinomycin D (7-AAD) Apoptosis Detection Kit (Vazyme, Nanjing, China). Flow cytometry was used to determine the percentages of apoptotic cells. Each assay contained three technical replicates and three biological replicates.

Total protein was extracted from the hippocampus and cortex of rats, SH-SY5Y cells, or HEK293 cells using RIPA Buffer (CWBio, Beijing, China) and then quantified using the bicinchoninic acid (BCA) Protein Assay Kit (Thermo Fisher, USA). The total protein was separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to nitrocellulose membranes (Millipore, Billerica, MS, USA). After being blocked with 5% skimmed milk for 1 h at 25 °C, the membranes were incubated with primary antibodies ( Table S1 ) at 4°C for 8 h. The membranes were then washed three times by Tris-Buffered Saline with Tween-20 (TBST) buffer and incubated with goat anti- (rabbit or mouse) secondary antibody ( Table S1 ) for 1 h at 25°C. A Fusion-FX6 imaging system (Vilber Lourmat, Marne-la-Vallée, France) was used for chemiluminescence color rendering and band analysis. Each analysis contained three biological replicates.

The tumor necrosis factor-alpha (TNF-α), interleukin-6 (IL-6), and caspase-3 levels in tissues and cells were measured using the corresponding ELISA kits (ImmunoWay Biotechnology Company, Plano, TX, USA) following the manufacturers’ instructions. Each analysis contained three technical replicates and three biological replicates.

The miR-152b-3p specific binding sites in the 3’-untranslated regions (3’-UTR) of CAMKIIA and the miR-193-3p binding sites in the 3’-UTR region of CALM were predicted using TargetScanHuman. About 150 bp before and after the binding site of CALM 3’-UTR and miR-193b-3p, CAMKIIA 3’-UTR and miR-152-3p were copied into a GP-miRGLO vector to construct luciferase wild-type (WT) and mutant (MUT) plasmids, in which the MUT plasmid only mutates the binding site. The wild-type (pGL-3’-UTR-CAMKIIA-WT and pGL-3’-UTR-CALM -WT) and mutant (pGL-3’-UTR-CAMKIIA-MUT and pGL-3’-UTR-CALM-MUT) plasmids were then transfected into HEK293 cells using the transfection methods described above. The miR-152-3p mimics, miR-193b-3p mimics, and NC were simultaneously co-transfected. The Dual-Luciferase Reporter Assay Kit (Vanzyme, Nanjing, China) was used 36 h post-transfection and the Renilla luciferase report system was used as a control. Each analysis contained three technical replicates and four biological replicates.

Statistical analyses were performed using SPSS software (Version 18.0, SPSS, Inc., Chicago, USA). One-way repeated measures analysis of variance (ANOVA) was used to compare the escape latency and swimming speed in the MWM test between groups. One-way ANOVA with Tukey’s post hoc test or Student’s unpaired t-test was used to analyze other data. Data are reported as the mean ± standard deviation (S.D.). Differences with a P-value < 0.05 were considered statistically significant.

The 2VO rats exhibited impaired learning compared with sham rats, as manifested by longer escape latency in the orientation navigation test ( Figure 1D , P < 0.01 vs. sham). Treatment with 40 mg/kg tilianin significantly reduced the escape latency of 2VO rats ( Figure 1D , P < 0.01 vs.2VO). However, there was no significant difference in swimming speed among the groups, indicating no impairment in motor ability ( Figure 1E ). In the probe trial, 2VO rats showed fewer crossings of the position where the previous platform was located and a shorter time spent in the quadrant without the previous platform ( Figures 1F, G , P < 0.01, P < 0.001 vs. sham). Tilianin treatment significantly improved memory, as demonstrated by an increased number of platform crossings and longer duration within the target quadrant compared with 2VO rats ( Figures 1F, G , P<0.05, P<0.01, respectively, vs. 2VO). It should be noted that 40 mg/kg tilianin did not significantly affect learning and memory in sham rats.

Nissl and TUNEL staining were used to evaluate histopathological changes in the brain. The number of Nissl bodies was greatly reduced in the hippocampus and cortex of 2VO rats ( Figures 1H, I , P < 0.01, P < 0.001 vs. sham). However, tilianin increased the number of Nissl bodies in these regions compared with 2VO rats ( Figures 1H, I , P < 0.05, P < 0.01 vs. 2VO). TUNEL staining showed that there was a significant increase in apoptosis in the 2VO model group ( Figures 1J, K , both P < 0.001 vs. sham) and a significant reduction of apoptosis in response to tilianin treatment ( Figures 1J, K , both P < 0.001 vs. 2VO).

Activation of glial cells is one of the major features of neuroinflammation, of which activation of microglia and astrocytes are typical symbols. Immunohistochemistry of IBA-1 and GFAP staining showed an increased microglial and astrocytic reaction in the 2VO model group ( Figures 1L–O , P < 0.001 vs. sham). Similarly, the activation of microglia and astrocytes was attenuated in response to tilianin treatment ( Figures 1L–O , P < 0.01, P < 0.001 vs. 2VO). Together, these results showed that tilianin effectively alleviated cognitive deficits, neuronal degeneration, and glial neuroinflammation in 2VO rats.

High-throughput sequencing of rat brains identified 13 upregulated miRNAs and 22 downregulated miRNAs in the 2VO model group (as compared to the sham group; Figure 2Ai ). Furthermore, there were 38 upregulated miRNAs and 18 downregulated miRNAs in the tilianin-treated 2VO group compared with the 2VO group ( Figure 2Aii ). Moreover, 203 upregulated mRNAs and 602 downregulated mRNAs were detected in the 2VO model group compared with the sham group ( Figure 2Bi ), while 704 upregulated mRNAs and 192 downregulated mRNAs were identified in the tilianin treatment group (compared with the 2VO group; Figure 2Bii ). Through analysis of the miRNAs described above, we screened 7 miRNAs ( Table S3 ) and 17 mRNAs ( Table S4 ) that showed opposite trends between the 2VO and tilianin-treated 2VO groups. These miRNAs and mRNAs may contribute to the molecular mechanism underlying the treatment effects of tilianin for VaD.

Using the miRDB, miRWalk, and TargetScan databases, 350 mRNAs were predicted to be the target genes of the candidate miRNAs. Fifteen common targets were obtained by merging the 17 screened potential mRNAs and predicted mRNAs, which were considered core targets closely related to tilianin functions ( Figure 2C ). A protein-protein interaction (PPI) network diagram was drawn to deeply analyze these core targets using the STRING database ( Figure 2D ). The network comprised 15 nodes (targets) and 34 edges (interactive relationships between targets). The key targets identified included CALM1, CAMK2A, CAMK2B, insulin-like growth factor 2 (IGF2), collagen α 1 (COL1A1), and excitatory amino acid transporter 1 (SLC1A3).

Next, GO and KEGG enrichments were used to analyze the biological functions and signaling pathways of the 15 core target genes. In GO terms, the biological functions focused mainly on vascular development, cardiovascular system development, vascular morphogenesis, cell adhesion regulation, etc. The cell component was concentrated primarily on the extracellular matrix, basement membrane, and contractile fibers containing collagen. The molecular functions focused on collagen binding, protein domain-specific binding, extracellular matrix structural components, and platelet-derived growth factor binding ( Figure 2E ). KEGG pathway analysis revealed that these core targets involved in the effects of tilianin were significantly enriched in some neurofunction-associated signaling pathways, namely the PI3K/AKT, Ca²⁺, neurotrophin, and MAPK signaling pathways ( Figure 2F ). Furthermore, a deep interaction network of dysregulated miRNAs and mRNAs revealed interactive relationships for miR-193b-3p, miR-152-3p, CaM, and CaMKIIα, highlighting them as potential targets for tilianin action and deserving of further investigation ( Figure 2G ).

After determining the involvement of miR-193b-3p and miR-152-3p in the effects of tilianin against VaD, we confirmed that their expression was significantly decreased in the cortex and hippocampus of 2VO rats ( Figures 3A–D , P<0.05, P<0.01 vs. sham, respectively), but increased in response to tilianin treatment ( Figures 3A–D , P<0.01, P<0.001 vs. 2VO, respectively). The expression of miR-193b-3p and miR-152-3p was also reduced in OGD-injured SH-SY5Y cells at different time points ( Figures 3E, F , all P<0.001 vs. 0 h), but increased by 10 μM and 30 μM tilianin treatment for 24 h ( Figures 3G, H , P<0.05, P<0.01 vs. OGD, respectively); notably, there were no expression changes in control cells treated with tilianin.

Subsequently, we investigated whether miR-193b-3p and miR-152-3p impacted VaD. As expected, overexpression of miR-193b-3p and miR-152-3p loaded by AAVs improved spatial learning and memory in 2VO rats ( Figures 4A, C, D , P<0.05, P<0.01 vs. 2VO+NCM, respectively). There was no significant influence of miR-193b-3p and miR-152-3p on learning and memory in sham rats compared with sham rats infused with NCMs ( Figures 4A, C, D ). Motor function was not affected by the various treatments, as reflected by similar swimming speeds ( Figure 4B ).

In terms of histopathological changes, overexpression of miR-193b-3p and miR-152-3p increased the percentage area of Nissl bodies in the cortex and hippocampus of 2VO rats, indicating attenuated neuronal loss ( Figures 4E, G , P<0.01, P<0.001 vs. 2VO+NCM, respectively). Similarly, TUNEL staining revealed significantly reduced apoptosis by miR-152-3p and miR-193b-3p mimics in 2VO rats ( Figures 4F, H , P<0.05, P<0.01, P<0.001 vs. 2VO+NC, respectively). These findings suggest that upregulation of miR-193b-3p and miR-152-3p alleviates neuronal degeneration and apoptosis in VaD model rats.

According to the RNA sequencing data and bioinformatics results, CaM and CaMKIIα are potential targets for the functions of miR-193b-3p and miR-152-3p in VaD. Subsequent analyses indicated that miR-193b-3p expression was negatively correlated with CaM expression, whereas miR-152-3p expression was negatively correlated with CaMKIIɑ expression, in the brains of 2VO rats ( Figures 5A, B , P<0.05, P<0.01, R^{2 =} 0.402, R^{2 =} 0.649). The online bioinformatics analyses identified the specific binding sites of miR-193b-3p and miR-152-3p as CaM mRNA 3’-UTR and CaMKIIɑ mRNA 3’-UTR, respectively ( Figures 5C, D ). These target sequences for binding are conserved in humans, rats, mice, chimpanzees, and other mammals, etc. Based on these results, plasmids containing wild-type and mutant-type binding sites were constructed ( Figures 5E, F ). Luciferase activity significantly decreased in cells co-transfected with CALM-wild-type plasmids and miR-193b-3p mimics ( Figures 5G , P<0.05 vs. NCM), as well as in cells co-transfected with CAMK2A-wild-type plasmids and miR-152-3p mimics ( Figure 5H , P<0.001 vs. NCM). In contrast, there was no change in luciferase activity between miR-193b-3p/miR-152-3p mimics and NCM in cells transfected with mutant plasmids. Moreover, CaM expression at the mRNA and protein level was significantly suppressed by miR-193b-3p mimics and elevated by miR-193b-3p inhibitors ( Figures 5I, K, M , P<0.05, P<0.01, P<0.001 vs. NCM/NCI). Similar results were observed for the effects of miR-152-3p mimics and inhibitors on CaMKIIɑ mRNA and protein expression levels ( Figures 5J, L, N , P<0.05, P<0.01 vs. NCM/NCI). Taken together, these findings indicate that miR-193b-3p and miR-152-3p directly bind to and regulate CaM and CaMKIIɑ expression.

Next, we explored whether the protective effects of miR-193b-3p and miR-152-3p in VaD are dependent on CaM and CaMKII targeting. The MWM results showed that miR-193b-3p- and miR-152-3p-mediated improvement in spatial learning ( Figure 6A , P<0.05, P<0.001 vs. 2VO+NCM) and memory ( Figures 6C, D , P<0.05, P<0.01 vs. 2VO+NCM) in 2VO rats was significantly blocked by overexpression of CaM and CaMKIIα ( Figures 6A, C, D , P<0.05, P<0.01, P<0.001 vs. 2VO+193M/2VO+152M). No rats showed motor disability, as indicated by no differences in swimming speed between the CaM and CaMKIIα overexpression groups and other treatment groups ( Figure 6B ).

Western blot results revealed that the expression of ox-CaMKII, p-CaMKII, and CaMKIIα, along with the ratio of p-CaMKII/CaMKIIα, in the rat cortex and hippocampus were significantly suppressed by miR-193b-3p and miR-152-3p ( Figures 7A–D , P<0.05, P<0.01, P<0.001 vs. 2VO+NC), and that such downregulation was reversed by CaM and CaMKIIα overexpression ( Figures 7A–D , P<0.05, P<0.01, P<0.001 vs. 2VO+193M/2VO+152M). Furthermore, CaM expression was suppressed by miR-193b-3p mimics, but unaffected by miR-152-3p mimics ( Figures 7A–D , P<0.05, P<0.01 vs. 2VO+NC).

Additional western blot and ELISA analyses revealed that the inhibition of apoptosis caused by miR-193b-3p and miR-152-3p was prevented by the corresponding CaM and CaMKIIα overexpression, as reflected by reduction of the increased Bcl-2/Bax ratio induced by miR-193b-3p and miR-152-3p in the cortex and hippocampus of 2VO rats and an increase in the decreased caspase-3 and PARP activity levels that had been induced by the two miRNAs ( Figures 7E–G, J–L , P<0.05, P<0.01, P<0.001 vs. 2VO+152M/2VO+193M). Overexpression of CaM and CaMKIIα also reversed the anti-inflammatory effects of miR-193b-3p and miR-152-3p in the cortex and hippocampus of 2VO rats, as demonstrated by elevation of miRNA-induced decreased levels of p-p38/p38, p-p65/p65, TNF-ɑ, and IL-6 ( Figures 7E, H–J, M, N , P<0.05, P<0.01 vs. 2VO+193M/2VO+152M).

In OGD-injured cells, miR-193b-3p mimics suppressed the expression of CaM, ox-CaMKII, p-CaMKII, and CaMKII, as well as the ratio of p-CaMKII/CaMKIIα, whereas CaM overexpression blocked these changes ( Figures S2A, B , P<0.05, P<0.01, P<0.001 vs. 193M). The expression of ox-CaMKII, p-CaMKII, and CaMKIIα, as well as the ratio of p-CaMKII/CaMKIIα, was decreased by miR-152-3p mimics ( Figures S2E, F , P<0.05, P<0.01, P<0.001 vs. NCM), but increased by co-transfection with miR-152-3p mimics and CaMKIIα ( Figures S2E, F , P<0.05, P<0.01, P<0.001 vs. 152M). Moreover, the suppressed levels of p-p38/p38, p-p65/p65, TNF-ɑ, and IL-6 due to the miR-193b-3p and miR-152-3p mimics were significantly elevated by corresponding CaM and CaMKIIα overexpression ( Figures S2A, C–E, G, H , P<0.05, P<0.01, P<0.001 vs. 193M/152M). Collectively, these results indicate that miR-193b-3p and miR-152-3p improved cognition in the VaD model rats by attenuating inflammation via CaM and CaMKII.

To verify the roles of miR-193b-3p and miR-152-3p in the anti-VaD effects of tilianin, miR-193b-3p and miR-152-3p were downregulated and behavior was then assessed. The MWM results revealed that tilianin treatment with NCI substantially improved learning ( Figure 8A , P<0.01 vs. 2VO+NCI) and memory ( Figures 8C, D , P<0.001 vs. 2VO+NCI) in 2VO rats. However, the miR-193b-3p or miR-152-3p inhibitor blocked tilianin-associated cognitive improvement ( Figures 8A, C, D , P<0.05, P<0.01 vs. 2VO+NCI+tilianin). Non-significant changes in swimming speed between groups indicated that miR-193b-3p and miR-152-3p inhibitors had minimal effect on motor motivation ( Figure 8B ).

In the histopathological analyses, Nissl staining revealed that the miR-193b-3p or miR-152-3p inhibitor decreased the percentage area of Nissl bodies ( Figures 8E, F , P<0.01 vs. 2VO+NCI+tilianin) in the cortex and hippocampus of 2VO rats, which had been increased by tilianin treatment. Similarly, the TUNEL staining results showed that the tilianin-associated reductions in apoptosis in 2VO rats were reversed by the miR-193b-3p and miR-152-3p inhibitor ( Figures 8G, H , P<0.01 vs. 2VO+NCI+tilianin). Together, these findings suggest that the protective effects of tilianin against cognitive decline and neurodegeneration rely on miR-193b-3p and miR-152-3p.

CaM expression in the cortex and hippocampus of 2VO rats was significantly decreased by tilianin; this effect was blocked by the miR-193b-3p inhibitor, but not the miR-152-3p inhibitor ( Figures 9A, B, G, H , P<0.05, P<0.01, P<0.001 vs. 2VO or 2VO+tilianin+NCI). The expression of ox-CaMKII, p-CaMKII, and CaMKIIα, as well as the p-CaMKII/CaMKIIα ratio, in these two brain regions in 2VO rats was significantly suppressed by tilianin ( Figures 9A, B, G, H , P<0.05, P<0.01, P<0.001 vs. 2VO); these suppressive effects were blocked by the miR-193b-3p and miR-152-3p inhibitor ( Figures 9A, B, G, H , P<0.05, P<0.01 vs. 2VO+tilianin+NCI). The NC inhibitors did not affect the beneficial effects of tilianin.

Furthermore, tilianin attenuated 2VO-induced apoptosis in rats ( Figures 9A, C, D, G, I, J , P<0.05, P<0.01, P<0.001 vs. 2VO), and these reductions were reversed by the miR-193b-3p and miR-152-3p inhibitors ( Figures 9A, C, D, G, I, J , P<0.05, P<0.01 vs. 2VO+tilianin+NCI). As expected, tilianin attenuated inflammatory reactions, including inhibition of p-p38/p38, p-p65/p65, TNF-ɑ, and IL-6 levels in 2VO rats ( Figures 9A, E, F, G, K, L , P<0.01, P<0.001 vs. 2VO). These effects were blocked by the miR-193b-3p inhibitor and miR-152-3p inhibitor ( Figures 9A, E–G, K, L , P<0.05, P<0.01, P<0.001 vs. 2VO+tilianin+NCI).

The occlusive mechanism of miR-193b-3p/CaM and miR-152-3p/CaMKII dysregulation on the protective effects of tilianin was further investigated using OGD-injured cells. When miR-193b-3p and miR-152-3p were downregulated, the protective effect of tilianin on the viability and apoptotic rates of OGD-injured cells was significantly inhibited ( Figures 10A–C , P<0.05, P<0.01 vs. OGD+tilianin+NCI). Notably, downregulation of miR-193b-3p and miR-152-3p inhibited the tilianin-induced expression decreases of CaM, ox-CaMKII, p-CaMKII, and CaMKII, along with the p-CaMKII/CaMKIIα ratio, in OGD-injured cells ( Figures 10D, E , P<0.05, P<0.01 vs. OGD+tilianin+NCI). The reverse effects were observed for the apoptotic and inflammatory indicators of tilianin-treated OGD-injured cells in the presence of miR-193b-3p and miR-152-3p inhibitors, as demonstrated by the decreased Bcl-2/Bax ratio, increased caspase-3 and PARP activity levels, and increased p-p38/p38, p-p65/p65, TNF-ɑ, and IL-6 levels ( Figures 10D, F–I , P<0.05, P<0.01, P<0.001 vs. OGD+tilianin+NCI). Upregulation of miR-193b-3p and miR-152-3p also enhanced the protective effects of tilianin on OGD-injured cells ( Figure 10J , P<0.05 vs. OGD+tilianin+NCM). When CaM and CaMKII were overexpressed, miR-193b-3p and miR-152-3p’s enhancement of tilianin’s neuroprotective effects in OGD-injured cells was suppressed ( Figure 10J , P<0.001 vs. OGD+tilianin+193M/152M). Similarly, CaM and CaMKII overexpression blocked the increased effects of miR-193b-3p and miR-152-3p on tilianin-mediated inflammation and apoptosis, as illustrated by activation of the p38 MAPK/NF-κB p65 and Bcl-2/Bax/caspase-3/PARP signaling cascades ( Figures 10K–O , P<0.05, P<0.01 and P<0.001, vs. OGD+tilianin+193M/152M). Taken together, these findings suggest that tilianin exerts protective effects against VaD by specifically acting on miR-193b-3p/CaM and miR-152-3p/CaMKII, jointly inhibiting the p38 MAPK/NF-κB p65 inflammatory and Bcl-2/Bax/caspase-3/PARP apoptotic pathways.