A novel miR-4536-3p inhibition ameliorates Alzheimer's disease by reducing Aβ accumulation and tau phosphorylation.

The human neuroblastoma cell line SH-SY5Y (RRID: CVCL_0019l ATCC ® CRL-2266) and HEK293T cell line (ATCC ® CRL-11268) were incubated in Dulbecco’s modified Eagle’s medium (DMEM, Cat. No. SH30243.01, Cytiva, Marlborough, MA, USA) supplemented with 10% fetal bovine serum (FBS, Cat. No. SV3020702, Cytiva), 1% penicillin–streptomycin (Cat. No. LS202-02, WELGENE, Gyeongsan-si, Republic of Korea), and 0.2% amphotericin B (Cat. No. 15290-018, Thermo Fisher Scientific, Waltham, MA, USA). The cells were maintained at 37℃ in a humidified environment with 5% CO₂ and 95% air [ 34 ]. Aβ 1−42 (Cat. No. AS-20276, ANASPEC, Fremont, CA, USA) was dissolved in 1% NH 4 OH and diluted to obtain a final stock concentration of 500 µM in the medium. This high concentration naturally promoted the formation of Aβ oligomers and aggregates, which was confirmed by transmission electron microscopy. For use as a therapeutic agent in the Aβ-induced AD model, miR-4536-3p-I (AGCACGUAUAUAGAUGGUGUA; Bioneer, Daejeon, Republic of Korea) was dissolved in ultra-pure distilled water (Invitrogen Co., Carlsbad, CA, USA) at concentrations of 0.1, 0.5, and 1 pM. SH-SY5Y cells were exposed to 500 nM Aβ 1−42 to establish an in vitro model of AD. After an hour of incubation, the cells were treated with 0.1, 0.5, and 1 pM miR-4536-3p-I. Subsequently, they were cultured in the presence or absence of Aβ 1−42 and miR-4536-3p-I for 24 h to conduct further experiments. miR-4536-3p mimic (miR-4536-3p-M, UCGUGCAUAUAUCUACCACAU; Bioneer), negative control inhibitor (NC-I; Bioneer), and negative control mimic (NC-M; Bioneer) were used for qPCR analysis and luciferase assay to confirm the DBN1 expression. For DBN1 overexpression, the coding sequence of DBN1 was cloned into the pcDNA3 vector (Invitrogen) to generate the pcDNA3-DBN1 construct. SH-SY5Y cells were first exposed to Aβ 1−42 for 1 h and then co-transfected with miR-4536-3p-M and pcDNA3-DBN1 to evaluate the rescue effect of DBN1 overexpression compared to cells treated with Aβ and miR-4536-3p-M alone. After 24 h of transfection, cells were harvested for quantitative real-time PCR (qPCR) and western blotting analyses. To investigate the expression pattern of miR-4536-3p, we utilized the miRBase ( https://www.mirbase.org/ ) and RNAcentral ( https://www.rnacentral.org/ ) databases. Tissue-specific expression was assessed using the Tissue Atlas Database ( https://ccb-compute2.cs.uni-saarland.de/mirnatissueatlas_2025 ). Target genes possessing complementary sequences with miR-4536-3p were identified using Targetscan ( https://www.targetscan.org/vert_80/ ), miRDB ( https://mirdb.org/ ), and miRTarbase ( https://mirtarbase.cuhk.edu.cn/ ). Specifically, genes from human, mouse, and rat that exhibited a higher degree of complementarity with miR-4536-3p compared to other miRNAs were selected. Among them, DBN1 , which contains a miR-4536-3p complementary region in its 3′-UTR in all three species, was selected as a potential target gene. An analysis of the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways and gene ontology (GO) related to miR-4536-3p was conducted using mirwalk ( http://mirwalk.umm.uni-heidelberg.de ) [ 35 , 36 ]. The pmirGLO Dual-Luciferase miRNA Target Expression Vector (Cat. No. E1330, Promega, Madison, WI, USA) was used to validate that miR-4536-3p could bind to the 3′-UTR of DBN1 [ 37 ]. The 3′-UTR of the predicted mRNA containing the hsa-miR-4536-3p target site was amplified and inserted into the pmirGLO vector [ 38 ]. A mutant version of the miR-4536-3p target site within the pmirGLO-DBN1 vector was amplified using an EZchange™ site-directed mutagenesis kit (Cat. No. EZ004S, Enzynomics, Daejeon, Republic of Korea). HEK293T cells were cotransfected with either wild-type or mutant pmirGLO-DBN1 plasmids and miR-4536-3p-M or NC-M using Lipofectamine 3000 (Cat. No. 2412049, Invitrogen). After 48 h, the cells were harvested, and the luciferase activity was measured. Relative luciferase activity was assessed using a dual-luciferase reporter assay system (Cat. No. E2940, Promega) and determined by calculating the ratio of firefly to Renilla luciferase signals. After the treatment with Aβ and miR-4536-3p-I, the total RNA was extracted from SH-SY5Y cells by using TRIzol (Cat. No. 9109, Takara Bio Co., Shimogyo-ku, Japan) in accordance with the previously established protocol [ 39 ]. Subsequently, cDNA was synthesized from the extracted RNA by using a PCR instrument (Takara Bio Co.) [ 40 ]. qPCR analyses were conducted to assess the expression of the target gene in SH-SY5Y cells treated with miR-4536-3p-I. Target genes were amplified using LightCycler 480 II (Roche Holding AG., Basel, Switzerland) and SYBR Green Premix Ex Taq (Cat. No. RR420, Takara Bio Co.) at an annealing temperature of 60℃. Three replicative reactions were performed for each sample. All primers (Table  1 ) were purchased from Macrogen (Seoul, Republic of Korea), CosmoGenetech (Seoul, Republic of Korea), and Bioneer. GAPDH was used as the endogenous control. Table 1 Sequence of qPCR primers Gene Forward (5′-3′) Reverse (5′-3′) Reference DBN1 GCCCCACCTGCTAACCAA GTGATTGACTGAAGTACCCCTCACT NM_001363541.2 [ 41 ] MAP2 CGGATCAACAGACAACATC CTGTGGCGGATGTTCTTC NM_001375534.1 [ 42 ] Tuj1 CGATGCCAAGAACATGATG CTCATCGACCTCCTTCATG AF 141349.1 [ 42 ] NEFL AGCCGTACTACTCGACCTCC GACTGGGCATCAACGATCCA NM_006158.5 [ 42 ] NEFM GCGCAAAGACTACCTGAAG GGCCTGGTGCATATTCTG NM_005382.2 [ 42 ] NEFH GTGGACCTGCAGAAGAAG CACCTCTTCCTGGTGGTG NM_021076.4 [ 42 ] GAPDH GACAGTCAGCCGCATCTTCT GCGCCCAATACGACCAAATC NM_002046.7 [ 42 ] Sequence of qPCR primers After the treatment with either Aβ or miR-4536-3p-I, apoptosis in SH-SY5Y cells was assessed through FACS analysis. After being washed with cold PBS, the cells were suspended in a 1X binding buffer and treated with FITC Annexin V and PI (Cat. No. 556547, BD Biosciences, Franklin Lakes, New Jersey, NJ, USA) in the dark for 15 min. They were analyzed using FACSCantoII (BD Biosciences), and the obtained results were processed in BD FACSDiva software (BD Biosciences). The 5xFAD transgenic mouse model (B6SJL) obtained from Jackson Laboratory (Bar Harbor, ME, USA) was used. These mice overexpress the human amyloid beta precursor protein 695 (APP) with three familial Alzheimer’s disease (FAD) mutations: Swedish (K670N, M671L), Florida (I716V), and London (V717I), along with the mutant presenilin 1 gene carrying two additional FAD mutations (M146L and L286V). Hemizygous transgenic 5xFAD mice were crossed with B6SJL F1/J mice (Jackson Laboratory) to maintain the mouse strain. For behavioral experiments, the mice were randomly divided into four groups: C57BL/6 (WT, n  = 14), 5xFAD transgenic mice (5xFAD, n  = 8), 5xFAD transgenic mice injected with NC-I (5xFAD + NC, n  = 8), and 5xFAD transgenic mice injected with miR-4536-3p-I (5xFAD + miR-4536-3p-I, n  = 8). For histological analysis including immunostaining and western blotting analysis, 4 to 6 mice per group were used, based on previously published protocols [ 38 , 43 – 45 ]. All experimental procedures were compliant with guidelines of the Institutional Animal Care and Use Committee (CNU IACUC-H-2024-39) at Chonnam National University Medical School [ 46 ]. All mice were housed under standard conditions of 22 ± 2℃ with 50% ± 10% relative humidity, with free access to food and water under a 12-hour light/dark cycle. No more than five mice were kept in one cage. Throughout the study, all subjects survived. To minimize stress caused by human interaction, mice were handled prior to cognitive testing as a conventional practice [ 47 , 48 ]. The mice were anesthetized with pentobarbital sodium (50 mg/kg; JW Pharmaceuticals, Seoul, Republic of Korea) and then positioned on a stereotaxic injector. miR-4536-3p-I (0.2 nmol) or NC (0.2 nmol) was injected into the hippocampal CA1 region (A/P, − 2 mm; M/L, ± 1.5 mm; D/V, − 1.5 mm) by using a Hamilton syringe (Cat. No. 788130, kd scientific, Seoul, Republic of Korea) at a concentration of 100 µM and an infusion rate of 0.5 µL/min. After surgery, the mice were allowed to recover for 3 days before they were subjected to behavioral testing. Then, they were sacrificed, and their brains were collected. The brain tissues were either fixed in 4% paraformaldehyde (PFA; Cat. No. SM-P01-100, GeneAll, Seoul, Republic of Korea) for histological analysis or frozen at − 80℃ for biochemical assays. The MWM was divided into four quadrants, and a hidden platform was submerged in one of them. Water was colored to obscure the platform, and the mice were trained to locate it, starting from the opposite quadrant. The time taken to reach the platform (escape latency) was recorded, and the maximum time limit was 60 s per trial. If a mouse failed to find the platform within this time, it was assigned a latency of 60 s. Regardless of the outcome, each mouse was allowed to remain on the platform for 15 s. Four training trials per mouse were conducted daily for 4 consecutive days, with 30-minute intervals between trials. A probe test was performed on the day following the final training session during which the platform was removed, and the time spent in each quadrant was measured to evaluate memory retention. Data such as frequency of platform crossing, time spent in each quadrant, and swimming speed were analyzed using EthoVision XT 17 software (Scitech Korea Inc., Seoul, Republic of Korea). A Cell Signaling Phospho Antibody Array (E-Biogen Inc., Seoul, Republic of Korea) was used to investigate the effect of miR-4536-3p-I on specific signaling pathways. This array includes 304 highly specific antibodies targeting phosphorylated proteins involved in 16 major cell signaling pathways, including PI3K/Akt, apoptosis, autophagy, ErbB, and JAK-STAT signaling [ 49 ]. The array is a high-throughput ELISA-based platform that enables simultaneous detection of multiple phospho-proteins from various sample types such as cell lysates, tissues, serum, and culture media. In this study, 3 × 10⁵ cells were seeded in a 6-well plate. After 24 h, the medium was replaced, and Aβ 1−42 (500 nM) was added to the cells. One hour later, miR-4536-3p-I (1 pM) was administered. After 24 h of treatment, the cells were collected as pellets without using trypsin to preserve protein integrity. Protein expression profiles were then compared between the Aβ-treated group and the miR-4536-3p-I-treated group. Protein-protein interaction analysis of the differentially expressed proteins identified in the antibody array was conducted using STRING ( https://string-db.org/ ). GO and KEGG pathway enrichment analyses, as well as disease association analyses, were conducted using EXDEGA (E-Biogen, Inc.) and DAVID Bioinformatics Resources ( https://davidbioinformatics.nih.gov/ ). In each group, the cells (3 × 10 5 cells/well) were treated with Aβ 1−42 and miR-4536-3p-I. They were harvested using a lysis buffer (1 M NaCl, 50 mg/mL aprotinin, 50 mg/mL leupeptin, 1 M Tris (pH 7.5), 10% Nonidet P-40, 100% glycerol, 0.1 M Na 3 VO 4 , 0.5 M EDTA, 0.2 M PMSF, and 1 M NaF). In parallel, the whole cerebral cortex and hippocampus were isolated from mouse brains and homogenized in the same lysis buffer. Protein concentrations were measured using a BCA protein assay kit (Cat. No. 23227, Thermo Fisher Scientific). The same amount of protein (15 µg) from distinct samples was separated using either 10% or 6% sodium dodecyl sulfate-polyacrylamide gel electrophoresis. The proteins were subsequently transferred to a polyvinylidene difluoride membrane (Cat. No. IPVH00010, Merck Millipore, Burlington, MA, USA). The primary and secondary antibodies are listed in Table  2 . β-actin was used as the primary endogenous loading control for western blotting. GAPDH served as an alternative loading control when either the target protein overlapped in molecular weight with β-actin or the β-actin expression level appeared inconsistent upon quantification. An Immobilon Crescendo Western HRP substrate (Cat. No. WBLUR0500, Merck Millipore) was used to visualize the immunoblotted bands, and quantitative analysis was conducted using Image J software. Table 2 Antibody list Antibody Company Catalog number Antibody ratio Bax Santa Cruz Biotechnology sc493 1:3000 Bcl-2 Santa Cruz Biotechnology sc492 1:500 Mcl-1 Cell Signaling Technology 94296 1:1000 Pro-caspase 9 Cell Signaling Technology 9508 1:1000 Pro-caspase 7 Cell Signaling Technology 12827 1:1000 Pro-caspase 3 Cell Signaling Technology 14220 1:1000 Pro-PARP Cell Signaling Technology 9542 1:1000 NeuN Sigma-Aldrich MAB377 1:2000 NFH Cell Signaling Technology 2836 1:1000 Tuj1 Cell Signaling Technology 5568 1:1000 Synaptophysin (SYP) Cell Signaling Technology 36406 1:3000 DBN1 Santa Cruz Biotechnology sc374269 1:1000 p-Akt Cell Signaling Technology 9271 1:1000 Akt Cell Signaling Technology 9272 1:1000 p-GSK3β Cell Signaling Technology 5558 1:3000 GSK3β Cell Signaling Technology 12456 1:3000 GAPDH Cell Signaling Technology 8884 1:3000 β actin Cell Signaling Technology 5125 1:3000 Horseradish peroxidase (HRP)-conjugated goat anti-rabbit IgG Cell Signaling Technology 7074 1:1000 HRP-conjugated goat anti-mouse IgG Cell Signaling Technology 7076 1:1000 Antibody list For the analysis of the protein expression in brain tissues, the mice were anesthetized with sodium pentobarbital and transcardially perfused with 4% PFA. Their brains were harvested, post-fixed in 4% PFA, and cryoprotected in 30% sucrose for at least 3 days. The tissues were embedded in optimal cutting temperature (OCT) compound (Cat. No. 3801480, Leica Biosystems, Nussloch, Germany) and sectioned into 10 μm slices using a DE/CM1860 cryostat (Leica Biosystems). For diaminobenzidine (DAB) staining, the brain sections were incubated with primary antibodies, namely anti-Aβ (Cat. No. sc28365, 1:1000, Santa Cruz Biotechnology), Bax (Cat. No. sc493, 1:1000, Santa Cruz Biotechnology), Bcl-2 (Cat. No. sc492, 1:1000, Santa Cruz Biotechnology), and NeuN (Cat. No. 24307, 1:1000, Cell Signaling Technology), diluted in 10% normal goat serum (NGS; Cat. No. 005-000-121, Jackson ImmunoResearch Labs, West Grove, PA, USA), for 1.5 h at room temperature. Subsequently, the sections were treated with biotinylated secondary antibodies (specific for rabbit or mouse IgG) and processed using the Vector Elite ABC kit (Cat. No. PK-6100, 1:100, Vector Laboratories, Newark, CA, USA). Signals were visualized with 0.03% DAB (Cat. No. ENZ-KIT159-0150, Enzo Life Sciences, Farmingdale, NY, USA), and the nuclei were counterstained with hematoxylin (Cat. No. S3309, Agilent, Santa Clara, CA, USA). For both representative images and quantitative analyses, we selected granular retrosplenial cortex and hippocampal regions that exhibited the most pronounced differences between WT and 5xFAD mice. The same subregions were consistently analyzed across all experimental groups to ensure accurate comparison. The immunoreactive area was quantified using Image J software. For immunofluorescence staining, the brain sections were incubated with primary antibodies, namely anti-NFH (Cat. No. 2836, 1:200, Cell Signaling Technology) and anti-MAP2 (Cat. No. 8707, 1:200, Cell Signaling Technology), diluted in 10% NGS. Alexa Fluor 488-conjugated goat anti-mouse (Cat. No. A11001, Invitrogen, 1:200) and Alexa Fluor 594-conjugated goat anti-rabbit (Cat. No. A11012, Invitrogen, 1:200) were used as secondary antibodies. The nuclei were counterstained with 4′,6-diamidino-2-phenylindole (DAPI, Cat. No. D1306, 1 µg/mL, Life Technologies Corporation, Carlsbad, CA, USA) for 30 min. Fluorescent signals were examined using a ZEISS Axio Vert.A1 light microscope (Carl Zeiss, Gottingen, Germany). The florescence intensity was quantified using image J software. Data were statistically analyzed using GraphPad Prism ® 10.5.0 (GraphPad Software Inc., San Diego, CA, USA) and shown as means ± standard error of the mean (SEM). Differences among the groups were assessed using one-way ANOVA followed by post-hoc Tukey’s tests. Differences between the two groups were examined using Student’s t-test. Significance levels were denoted as follows: * p  < 0.05, ** p  < 0.01, and *** p  < 0.001 compared with the control or WT group; # p  < 0.05, ## p  < 0.01, and ### p  < 0.001 compared with the Aβ group or 5xFAD + NC group.

We performed a microarray analysis to identify the potential miRNA candidates by comparing their expression levels between hADSC and NI-hADSC (Fig. S1 ) [ 50 ]. Cells were isolated from three independent human donors and individually characterized. Among the differentially expressed miRNAs, miR-647, miR-4328, and miR-4536-3p were consistently upregulated in NI-hADSCs compared to hADSCs across all three samples. miR-647 has been previously reported to modulate AD progression [ 51 ], and miR-4328 has been implicated in apoptosis and lung adenocarcinoma [ 52 , 53 ]. Interestingly, miR-4536-3p has not yet been investigated in any biological context. Based on this, we selected miR-4536-3p as a novel candidate for further investigation into its functional role in AD. To evaluate the efficiency of miR-4536-3p modulation, we performed miRNA PCR on HEK293T cells treated with miR-4536-3p-I or miR-4536-3p-M at 1 pM, 100 pM, and 25 nM. The treatment with 1 pM (0.026 ± 0.375) and 100 pM (0.080 ± 0.197) miR-4536-3p-I significantly reduced the miR-4536-3p expression (Fig. S2 A). In contrast, miR-4536-3p-M treatment led to a marked, dose-dependent increase in miR-4536-3p expression at 100 pM (128.890 ± 0.092) and 25 nM (300,429.400 ± 0.044, Fig. S2 B). Based on the dose-response analysis, treatment with the miR-4536-3p-I significantly reduced miR-4536-3p expression even at the lowest tested concentration of 1 pM. Therefore, 1 pM was selected as the working concentration for subsequent in vitro experiments to achieve effective inhibition while minimizing potential off-target effects. Additionally, Fig. S2 C shows the expression of miR-4536-3p across various human tissues based on the Tissue Atlas Database, indicating detectable expression in brain samples. This observation provides a rationale for investigating its role in neurodegenerative disorders such as AD. Subsequently, we searched the target genes of miR-4536-3p following Targetgene, miRDB, and miRTarbase. Among the predicted targets, four genes— PFKP , NUPL2 , DBN1 , and HNRNPU —were commonly identified across all three platforms (Fig.  1 A). Of these, DBN1 was selected for further analysis due to its known downregulation in AD and potential relevance as a miR-4536-3p target. qPCR analysis confirmed that the inhibition of miR-4536-3p induced the mRNA expression of DBN1, whereas miR-4536-3p-M decreased the mRNA expression of DBN1 compared with that of NC in SH-SY5Y cells (Fig.  1 C). In contrast, the mRNA expression levels of Wnt4 and Olig3 , candidate target genes of miR-4536-3p, were not significantly affected by either miR-4536-3p inhibition (miR-4536-3p-I) or overexpression (miR-4536-3p-M) in SH-SY5Y cells (Fig. S3 ). Figure  1 B shows the sequence alignment of miR-4536-3p and 3′-UTR of DBN1. Additionally, the relative luciferase activity decreased with miR-4536-3p-M and pmirGLO-DBN1 WT in HEK293T cells; conversely, it remained unchanged in the mutant (Fig.  1 D). Western blotting analysis revealed that the DBN1 protein expression increased in the brain of the 5xFAD mice injected with miR-4536-3p-I compared with that in the untreated 5xFAD mice and the 5xFAD mice injected with NC (Fig.  1 E–F). Therefore, miR-4536-3p could be a target 3′-UTR of DBN1 and inhibited the mRNA expression and translation of DBN1. Fig. 1 Upregulation of DBN1 following miR-4536-3p-I treatment. ( A ) The Venn diagram displays the number of the predicted target genes of miR-4536-3p based on analysis using Targetscan, miRDB, and miRTarbase software. ( B ) Sequence alignment illustrates the predicted binding sites between miR-4536-3p and the 3ʹ-UTR of the mRNA of DBN1, with complementary sequences highlighted in red. ( C ) qPCR analysis demonstrated a significant increase in the DBN1 expression following miR-4536-3p-I treatment, where mRNA expression was normalized to the NC group and GAPDH was used as the endogenous control ( n  = 3 per group). ( D ) Relative luciferase activity was measured in the cells transfected with either NC-M or miR-4536-3p-M, along with pmiGlo-DBN1 WT or mutant constructs ( n  = 4–6 per group). The activity was reduced in the WT but not in the mutant construct. ( E - F ) Western blotting analysis showed an increase in the DBN1 protein expression in the 5xFAD + miR-4536-3p-I group compared with that in the 5xFAD or 5xFAD + NC groups ( n  = 4–6 per group). GAPDH was used as the endogenous control. Results represent means ± SEMs. Statistical significance is indicated as follows: ** p  < 0.01 and *** p  < 0.001 compared with the NC group. ## p  < 0.01 compared with the 5xFAD or 5xFAD + NC group Upregulation of DBN1 following miR-4536-3p-I treatment. ( A ) The Venn diagram displays the number of the predicted target genes of miR-4536-3p based on analysis using Targetscan, miRDB, and miRTarbase software. ( B ) Sequence alignment illustrates the predicted binding sites between miR-4536-3p and the 3ʹ-UTR of the mRNA of DBN1, with complementary sequences highlighted in red. ( C ) qPCR analysis demonstrated a significant increase in the DBN1 expression following miR-4536-3p-I treatment, where mRNA expression was normalized to the NC group and GAPDH was used as the endogenous control ( n  = 3 per group). ( D ) Relative luciferase activity was measured in the cells transfected with either NC-M or miR-4536-3p-M, along with pmiGlo-DBN1 WT or mutant constructs ( n  = 4–6 per group). The activity was reduced in the WT but not in the mutant construct. ( E - F ) Western blotting analysis showed an increase in the DBN1 protein expression in the 5xFAD + miR-4536-3p-I group compared with that in the 5xFAD or 5xFAD + NC groups ( n  = 4–6 per group). GAPDH was used as the endogenous control. Results represent means ± SEMs. Statistical significance is indicated as follows: ** p  < 0.01 and *** p  < 0.001 compared with the NC group. ## p  < 0.01 compared with the 5xFAD or 5xFAD + NC group We performed an antibody array to compare the protein expression between Aβ-treated cells and Aβ-treated cells with the miR-4536-3p-I. We selected specific proteins that showed significant expression changes in Aβ-treated cells with miR-4536-3p-I compared with Aβ-treated cells alone for biological analysis using Exdega, DAVID, and STRING software (fold change ≥ 1.2). Figure  2 A highlights the top 15 diseases, including AD, based on the proteins identified from the antibody array. Figure  2 B illustrates the AD-related protein, with line thickness indicating the strength of the interaction. KEGG pathway and GO analysis further identified the top 15 pathways and mechanisms related to these specific proteins (Fig.  2 C–F). Interestingly, the PI3K-Akt signaling pathway and apoptotic processes were implicated in miR-4536-3p-I-treated AD cells in vitro. These findings suggested that the miR-4536-3p-I might influence apoptosis and contribute to AD pathology. Fig. 2 Protein expression profiling between the Aβ the Aβ with miR-4536-3p-I ( A ) Graph illustrates the enriched disease processes associated with differentially expressed proteins identified from the antibody array. Red bars indicate the number of upregulated proteins linked to specific diseases, and blue bars represent the downregulated proteins. The black dotted line indicates the statistical significance of these enrichments, represented as–log 10 (p-value). ( B ) A network diagram displays the interactions among proteins specifically related to AD. KEGG pathway analysis ( C ) and GO analysis ( D - F ) were conducted Protein expression profiling between the Aβ the Aβ with miR-4536-3p-I ( A ) Graph illustrates the enriched disease processes associated with differentially expressed proteins identified from the antibody array. Red bars indicate the number of upregulated proteins linked to specific diseases, and blue bars represent the downregulated proteins. The black dotted line indicates the statistical significance of these enrichments, represented as–log 10 (p-value). ( B ) A network diagram displays the interactions among proteins specifically related to AD. KEGG pathway analysis ( C ) and GO analysis ( D - F ) were conducted To demonstrate the effect of miR-4536-3p-I, we investigated its role in apoptosis and cell death. We found a significant increase in early apoptosis, late apoptosis, and total apoptosis (both early apoptosis and late apoptosis) in the Aβ-treated group compared with those in the control group. Notably, the percentage of late apoptosis in Aβ + miR-4536-3p-I significantly decreased compared with that in the Aβ-treated group regardless of the inhibitor concentration (Fig.  3 A–B). Similarly, the percentages of early apoptosis and total apoptosis considerably decreased at certain concentrations of miR-4536-3p-I compared with those in the Aβ-treated group. These results suggested that miR-4536-3p-I might effectively prevent Aβ-induced apoptosis. Western blotting analysis was performed to examine the expression of apoptotic markers in the in vitro AD model treated with miR-4536-3p-I. The results showed that miR-4536-3p-I (1 pM) increased the expression of Mcl-1, an anti-apoptotic marker, compared with that in the Aβ-treated group. The expression of Bcl-2, another anti-apoptotic marker, also increased with miR-4536-3p-I treatment, but the change was not statistically significant (Fig.  3 C–D). Conversely, the pro-apoptotic marker Bax indicated an increased expression following Aβ exposure. Furthermore, the levels of pro-caspase 9, pro-caspase 7, pro-caspase 3, and pro-PARP, which typically increase as the levels of their cleaved forms decrease, were upregulated after the treatment with a specific concentration of miR-4536-3p-I (Fig.  3 E–F). All uncropped gel images were shown in the supplementary data (Fig. S4 ). Therefore, miR-4536-3p-I exerted a protective anti-apoptotic effect on the in vitro AD model. Fig. 3 Inhibition of apoptotic processes using miR-4536-3p-I. ( A ) Flow cytometry images illustrated the effect of miR-4536-3p-I on the Aβ-induced apoptosis of SH-SY5Y cells. ( B ) Quantification of the apoptosis of Aβ-treated cells following miR-4536-3p-I treatment revealed a significant decrease in early apoptosis, late apoptosis, and total apoptosis compared with the Aβ-treated group ( n  = 3 per group). Western blotting analysis was used to examine the protein expression levels of key regulators such as Bcl-2, Mcl-1, and Bax ( C - D ) in the mitochondrial apoptosis pathway and the downstream apoptosis pathway, which includes pro-cas-9/-7/-3 ( E - F ) ( n  = 4–6 per group). Protein expression levels were normalized to the control group, and β-actin was used as the internal control. Results represent means ± SEMs. Statistical significance is indicated as ### p  < 0.001, ## p  < 0.01, and # p  < 0.05 compared with the Aβ-treated group Inhibition of apoptotic processes using miR-4536-3p-I. ( A ) Flow cytometry images illustrated the effect of miR-4536-3p-I on the Aβ-induced apoptosis of SH-SY5Y cells. ( B ) Quantification of the apoptosis of Aβ-treated cells following miR-4536-3p-I treatment revealed a significant decrease in early apoptosis, late apoptosis, and total apoptosis compared with the Aβ-treated group ( n  = 3 per group). Western blotting analysis was used to examine the protein expression levels of key regulators such as Bcl-2, Mcl-1, and Bax ( C - D ) in the mitochondrial apoptosis pathway and the downstream apoptosis pathway, which includes pro-cas-9/-7/-3 ( E - F ) ( n  = 4–6 per group). Protein expression levels were normalized to the control group, and β-actin was used as the internal control. Results represent means ± SEMs. Statistical significance is indicated as ### p  < 0.001, ## p  < 0.01, and # p  < 0.05 compared with the Aβ-treated group To confirm the expression of neuronal markers after Aβ toxicity, we conducted western blotting and qPCR analyses. After the treatment with miR-4536-3p-I after exposure to Aβ, we observed that the protein expression levels of NeuN, NFH, Tuj1, and SYP were upregulated compared with the levels detected after the Aβ treatment (Fig.  4 A–B). All uncropped gel images were shown in the supplementary data (Fig. S5 ). Regardless of concentration, miR-4536-3p-I treatment promoted neuronal markers such as MAP2 , Tuj1 , NEFM , and NEFL compared with those in the Aβ-treated group ( Fig.  4 C–F ) . Therefore, miR-4536-3p-I exhibited a neuroprotective effect on cell death induced by Aβ. Fig. 4 Neuroprotective effect of miR-4536-3p-I on Aβ-induced cell death. ( A - B ) The protein expression of neuronal markers (NeuN, NFH, Tuj1, and SYP) was examined by western blotting analysis ( n  = 4–6 per group). β-actin was used as the endogenous control. The gene expression levels of MAP2 ( C ), Tuj1 ( D ), NEFM ( E ), and NEFL ( F ), which are neuronal markers, were examined via qPCR analysis ( n  = 3 per group). GAPDH was used as the endogenous control. After the treatment with miR-4536-3p-I, the expression of neuronal markers significantly increased in Aβ-induced cells. Results represent means ± SEMs. Statistical significance is indicated as ### p  < 0.001, ## p  < 0.01, and # p  < 0.05 compared with the Aβ-treated group Neuroprotective effect of miR-4536-3p-I on Aβ-induced cell death. ( A - B ) The protein expression of neuronal markers (NeuN, NFH, Tuj1, and SYP) was examined by western blotting analysis ( n  = 4–6 per group). β-actin was used as the endogenous control. The gene expression levels of MAP2 ( C ), Tuj1 ( D ), NEFM ( E ), and NEFL ( F ), which are neuronal markers, were examined via qPCR analysis ( n  = 3 per group). GAPDH was used as the endogenous control. After the treatment with miR-4536-3p-I, the expression of neuronal markers significantly increased in Aβ-induced cells. Results represent means ± SEMs. Statistical significance is indicated as ### p  < 0.001, ## p  < 0.01, and # p  < 0.05 compared with the Aβ-treated group To investigate whether the therapeutic effect of miR-4536-3p inhibition is mediated by DBN1, we overexpressed DBN1 in SH-SY5Y cells co-treated with Aβ and the miR-4536-3p-M. qPCR analysis revealed that miR-4536-3p mimic treatment significantly reduced DBN1 mRNA levels in Aβ-treated cells, whereas DBN1 overexpression markedly increased DBN1 expression (Fig.  5 A–B). Consistently, western blotting analysis confirmed that DBN1 protein levels were significantly upregulated by DBN1 overexpression compared to both the Aβ group and the Aβ + miR-4536-3p-M group (Fig.  5 C–D). Next, qPCR analysis showed that the expression levels of neuronal markers, including MAP2, NEFH, and Tuj1, were significantly decreased by miR-4536-3p-M treatment in Aβ-treated cells (Fig.  5 E and G, and 5 I). However, DBN1 overexpression restored the expression of these neuronal markers (Fig.  5 F and H, and 5 J), suggesting that the neuroprotective effects of miR-4536-3p inhibition are mediated in a DBN1-dependent manner. Furthermore, western blotting indicated that the miR-4536-3p-M increased Bax expression, exacerbating apoptosis under Aβ toxicity, while DBN1 overexpression significantly downregulated Bax levels compared to the Aβ + miR-4536-3p-M group (Fig.  5 K–L). Collectively, these findings demonstrate that DBN1 is a key modulator mediating the neuroprotective and anti-apoptotic effects of miR-4536-3p inhibition in an in vitro AD model. Fig. 5 Therapeutic effects of DBN1 in an in vitro AD model. ( A – B ) Relative mRNA levels of DBN1 were decreased by miR-4536-3p-M treatment in Aβ-treated cells and restored by DBN1 overexpression (qPCR; n  = 3 per group). ( C – H ) Relative mRNA expression of neuronal markers was significantly downregulated by miR-4536-3p-M treatment in Aβ-treated cells, while DBN1 overexpression restored their expression (qPCR; n  = 3 per group). GAPDH was used as the endogenous control for qPCR analysis. ( I – J ) Representative western blotting analysis and quantification showing increased DBN1 protein levels after DBN1 overexpression compared to the Aβ group and the Aβ + miR-4536-3p-M group (western blotting analysis; n  = 4–6 per group). ( K – L ) Representative Western blot images and quantification showing that Bax expression was increased by miR-4536-3p-M treatment in Aβ-treated cells and reduced by DBN1 overexpression (western blotting analysis; n  = 4–6 per group). β-actin was used as the endogenous control for western blotting analysis. Results represent means ± SEMs. Statistical significance is indicated as ### p  < 0.001, ## p  < 0.01, and # p  < 0.05 compared with the Aβ-treated group; &&& p  < 0.001 and && p  < 0.01 compared with the Aβ + miR-4536-3p-M group Therapeutic effects of DBN1 in an in vitro AD model. ( A – B ) Relative mRNA levels of DBN1 were decreased by miR-4536-3p-M treatment in Aβ-treated cells and restored by DBN1 overexpression (qPCR; n  = 3 per group). ( C – H ) Relative mRNA expression of neuronal markers was significantly downregulated by miR-4536-3p-M treatment in Aβ-treated cells, while DBN1 overexpression restored their expression (qPCR; n  = 3 per group). GAPDH was used as the endogenous control for qPCR analysis. ( I – J ) Representative western blotting analysis and quantification showing increased DBN1 protein levels after DBN1 overexpression compared to the Aβ group and the Aβ + miR-4536-3p-M group (western blotting analysis; n  = 4–6 per group). ( K – L ) Representative Western blot images and quantification showing that Bax expression was increased by miR-4536-3p-M treatment in Aβ-treated cells and reduced by DBN1 overexpression (western blotting analysis; n  = 4–6 per group). β-actin was used as the endogenous control for western blotting analysis. Results represent means ± SEMs. Statistical significance is indicated as ### p  < 0.001, ## p  < 0.01, and # p  < 0.05 compared with the Aβ-treated group; &&& p  < 0.001 and && p  < 0.01 compared with the Aβ + miR-4536-3p-M group On the basis of our previous findings, we investigated whether miR-4536-3p-I could elicit therapeutic effects on an in vivo AD model. To identify the appropriate age of 5xFAD mice for studying AD-related pathological features, including cognitive deficits, Aβ deposition, apoptosis, and neuronal cell loss, we first conducted the MWM test by using 3-month-old WT and 5xFAD mice. Although the 5xFAD mice exhibited trends of increased escape latency, reduced time spent in the target quadrant, and fewer platform crossings compared with the WT mice, these differences were not statistically significant (Fig. S6 ). Considering these findings, we selected 4-month-old 5xFAD mice for further experiments to evaluate the cognitive effects of the miR-4536-3p-I treatment. Either miR-4536-3p-I or NC was stereotaxically injected into the hippocampus, and cognitive function was assessed using the MWM for 5 consecutive days (Fig.  6 A). To avoid acute stress responses and ensure stabilization of memory-related molecular changes, mice were sacrificed 3 days after completing the MWM test for histological analysis. To determine the effective concentration of miR-4536-3p-I in the in vivo AD model, we administered 50 µM and 100 µM miR-4536-3p-I into the hippocampus of 5xFAD mice, based on previous studies [ 43 , 44 ], and evaluated cognitive function using the MWM test. Both the 50 µM and 100 µM miR-4536-3p-I groups showed significantly reduced escape latency on day 4 compared to the 5xFAD + NC group (Fig. S7 A). Notably, the 100 µM miR-4536-3p-I group exhibited a significant reduction in escape latency beginning on day 3, indicating superior cognitive improvement compared to the 50 µM group. Swimming speed did not differ significantly among the groups (Fig. S7 B). Based on these findings, we selected 100 µM as the optimal concentration of miR-4536-3p-I for evaluating its therapeutic effects on cognitive impairment. Using this concentration, we further examined its effects on spatial learning and memory in 4-month-old 5xFAD mice. As shown in Fig.  6 B, inhibition of miR-4536-3p significantly improved spatial learning in 4-month-old 5xFAD mice compared with those injected with the NC. However, no significant difference was observed in the swimming speed between the groups (Fig.  6 D). In the probe trial, mice treated with the miR-4536-3p-I spent more time in the target quadrant and exhibited more platform crossings compared with the NC-treated 5xFAD mice (Fig.  6 C, E). Additionally, miR-4536-3p-I-treated 5xFAD mice exhibited more direct and efficient swimming paths to the platform, in contrast to the more circuitous routes observed in the NC-treated group (Fig.  6 F). Furthermore, the 5xFAD + miR-4536-3p-I group spent significantly more time around the platform during the probe trial compared with both the 5xFAD and 5xFAD + NC groups (Fig.  6 G). These results demonstrated that the inhibition of miR-4536-3p could mitigate cognitive deficits in AD mice by enhancing spatial learning and hippocampus-dependent memory. Fig. 6 Inhibition of miR-4536-3p improves cognitive function in AD mice. ( A ) Scheme explaining the in vivo experimental procedures in 4-month 5xFAD mice. MWM results showed in ( B - G ). Following the MWM, the latency to find the platform ( B ), duration within the target quadrant ( C ), swimming speed ( D ), and number of crossing the platform ( E ) were calculated among the groups. ( F ) Representative images of tracings showing the route of the mice finding the platform in the probe test. ( G ). Representative heatmap images indicating the routes taken by mice to find the platform in the probe test. Areas in red indicate locations where the mice spent more time than in other areas. Results represent means ± SEMs ( n  = 8–12 per group). Statistical significance is indicated as follows: ** p  < 0.01 and *** p  < 0.001 compared with the WT group. ### p  < 0.001 compared with the 5xFAD + NC group Inhibition of miR-4536-3p improves cognitive function in AD mice. ( A ) Scheme explaining the in vivo experimental procedures in 4-month 5xFAD mice. MWM results showed in ( B - G ). Following the MWM, the latency to find the platform ( B ), duration within the target quadrant ( C ), swimming speed ( D ), and number of crossing the platform ( E ) were calculated among the groups. ( F ) Representative images of tracings showing the route of the mice finding the platform in the probe test. ( G ). Representative heatmap images indicating the routes taken by mice to find the platform in the probe test. Areas in red indicate locations where the mice spent more time than in other areas. Results represent means ± SEMs ( n  = 8–12 per group). Statistical significance is indicated as follows: ** p  < 0.01 and *** p  < 0.001 compared with the WT group. ### p  < 0.001 compared with the 5xFAD + NC group Histological and immunofluorescence analyses revealed that miR-4536-3p inhibition significantly reduced apoptosis and neuronal loss in the 5xFAD mouse model of AD. The pro-apoptotic marker Bax was upregulated in the cerebral cortex and CA2 region of the hippocampus of 5xFAD and 5xFAD + NC groups compared with those of the WT group but was markedly downregulated after the treatment with the miR-4536-3p-I (Fig.  7 A–B). Conversely, the expression of the anti-apoptotic marker Bcl-2, which was decreased in the cerebral cortex and CA1 region of the hippocampus of 5xFAD and 5xFAD + NC groups, notably increased upon miR-4536-3p-I treatment (Fig.  7 C–D). NeuN, a neuronal marker, was significantly reduced in the cerebral cortex and CA2 region of the hippocampus of 5xFAD and 5xFAD + NC groups compared with that in the WT group, whereas miR-4536-3p-I treatment restored the NeuN expression to higher levels relative to the 5xFAD + NC group (Fig.  7 E–F). Similarly, the neuronal markers MAP2 and NFH were significantly decreased in the cerebral cortex and hippocampus of the 5xFAD and 5xFAD + NC groups compared with those in the WT group (Fig.  7 G–L). However, their expression was markedly increased in the 5xFAD + miR-4536-3p-I group compared with that in the 5xFAD + NC group (Fig.  7 G–L). Low-magnification images showing the overall histological features of both the cerebral cortex and hippocampus are provided in Fig. S8 . These results suggested that miR-4536-3p-I exerted therapeutic effects on 5xFAD mice by attenuating apoptosis and neuronal degeneration. Fig. 7 miR-4536-3p-I attenuates apoptosis and neurodegeneration in the cerebral cortex and hippocampus of 5xFAD mice. Representative images and quantitative analysis of Bax ( A - B ), Bcl-2 ( C - D ), and NeuN ( E - F ) in the cerebral cortex and hippocampus of the mouse groups. The scale bar indicates 200 μm. Representative images and quantitative analysis of NFH (green) and MAP2 (red) in the cerebral cortex ( G - I ) and hippocampus ( J - L ) of the mouse groups. The scale bar indicates 100 μm. Results represent means ± SEMs ( n  = 4 per group). Statistical significance is indicated as follows: ** p  < 0.01 and *** p  < 0.001 compared with the WT group. ## p  < 0.01 and ### p  < 0.001 compared with the 5xFAD + NC group miR-4536-3p-I attenuates apoptosis and neurodegeneration in the cerebral cortex and hippocampus of 5xFAD mice. Representative images and quantitative analysis of Bax ( A - B ), Bcl-2 ( C - D ), and NeuN ( E - F ) in the cerebral cortex and hippocampus of the mouse groups. The scale bar indicates 200 μm. Representative images and quantitative analysis of NFH (green) and MAP2 (red) in the cerebral cortex ( G - I ) and hippocampus ( J - L ) of the mouse groups. The scale bar indicates 100 μm. Results represent means ± SEMs ( n  = 4 per group). Statistical significance is indicated as follows: ** p  < 0.01 and *** p  < 0.001 compared with the WT group. ## p  < 0.01 and ### p  < 0.001 compared with the 5xFAD + NC group To assess the effect of miR-4536-3p inhibition on AD pathology, we performed IHC staining and western blotting analysis in 5xFAD mice. Aβ plaques, a hallmark of AD pathology, significantly accumulated in the brains of 5xFAD mice. However, Aβ plaques in the cerebral cortex and CA1 region of the hippocampus were markedly reduced by the treatment with miR-4536-3p-I compared with those in the 5xFAD + NC group (Fig.  8 A–B). Low-magnification images showing the overall histological features of both the cerebral cortex and hippocampus are provided in Fig. S8 . In the cerebral cortex, the phosphorylated Tau-to-Tau (p-Tau/Tau) ratio, a key marker of tau pathology, was significantly decreased in the 5xFAD + miR-4536-3p-I group compared with that in the 5xFAD and 5xFAD + NC groups (Fig.  8 C–D). Furthermore, the ratio of phosphorylated GSK3β-to-GSK3β (p-GSK3β/GSK3β), which indicated GSK3β inactivation, was significantly increased in the same group (Fig.  8 C and E). Similar trends were observed in the hippocampus, with reduced p-Tau/Tau levels and increased p-GSK3β/GSK3β levels after the miR-4536-3p-I treatment (Fig.  8 F–H). All uncropped gel images were shown in the supplementary data (Fig. S9 ). These findings indicated that miR-4536-3p inhibition mitigated AD progression by reducing Aβ deposition and tau phosphorylation and by promoting GSK3β inactivation. Fig. 8 miR-4536-3p-I recovers AD progress in the cerebral cortex and hippocampus of 5xFAD mice. Representative images ( A ) and quantitative analysis ( B ) of Aβ in the cerebral cortex and hippocampus of 5xFAD mice following miR-4536-3p-I or NC treatment ( n  = 4 per group). The scale bar indicates 200 μm. ( C - E ) Western blotting analysis of p-Tau, Tau, p-GSK3β, and GSK3β expression in the cerebral cortex of WT or 5xFAD mice following miR-4536-3p-I treatment. ( F - H ) Western blotting analysis of p-Tau, Tau, p-GSK3β, and GSK3β expression in the hippocampus of WT or 5xFAD mice following miR-4536-3p-I treatment. GAPDH was used as the endogenous control. Results represent means ± SEMs ( n  = 4–6 per group). Statistical significance is indicated as follows: # p  < 0.05, ## p  < 0.01, and ### p  < 0.001 compared with the 5xFAD or 5xFAD + NC group miR-4536-3p-I recovers AD progress in the cerebral cortex and hippocampus of 5xFAD mice. Representative images ( A ) and quantitative analysis ( B ) of Aβ in the cerebral cortex and hippocampus of 5xFAD mice following miR-4536-3p-I or NC treatment ( n  = 4 per group). The scale bar indicates 200 μm. ( C - E ) Western blotting analysis of p-Tau, Tau, p-GSK3β, and GSK3β expression in the cerebral cortex of WT or 5xFAD mice following miR-4536-3p-I treatment. ( F - H ) Western blotting analysis of p-Tau, Tau, p-GSK3β, and GSK3β expression in the hippocampus of WT or 5xFAD mice following miR-4536-3p-I treatment. GAPDH was used as the endogenous control. Results represent means ± SEMs ( n  = 4–6 per group). Statistical significance is indicated as follows: # p  < 0.05, ## p  < 0.01, and ### p  < 0.001 compared with the 5xFAD or 5xFAD + NC group To investigate the molecular mechanisms through which miR-4536-3p-I mitigates AD-like symptoms, we performed an antibody array analysis on Aβ-treated cells with and without miR-4536-3p-I. The resulting heat map showed a reduction in apoptotic markers, including caspase 9, Bax, and caspase 3, in the Aβ + miR-4536-3p-I group compared with those in the Aβ group (Fig.  9 A). Furthermore, the key proteins in the PI3K/Akt signaling pathway—, namely, PDK1, p-Akt/Akt, and p-mTOR/mTOR—increased in the Aβ + miR-4536-3p-I group compared with those in the Aβ group. The network map in Fig.  9 B illustrates the interactions within the PI3K/Akt pathway, with line thickness representing interaction strength. Following in vitro results, we further assessed the PI3K/Akt pathway activation in the cerebral cortex and hippocampus of the 5xFAD mice treated with either miR-4536-3p-I or NC. Consistent with the in vitro findings, the expression of PI3K and p-Akt/Akt significantly increased in the cerebral cortex of miR-4536-3p-I-treated 5xFAD mice compared with those in the 5xFAD mice (Fig.  9 C–E). Similar effects were observed in the hippocampus; specifically, PI3K and p-Akt/Akt levels significantly increased in the 5xFAD + miR-4536-3p-I group compared with those in the 5xFAD or 5xFAD + NC group (Fig.  9 F–H). All uncropped gel images were shown in the supplementary data (Fig. S10 ). These results suggested that miR-4536-3p-I activated the PI3K/Akt signaling pathway, potentially attenuating AD-like symptoms and apoptosis in in vitro and in vivo AD models Fig. 9 Expression of PI3K/Akt/GSK3β signaling pathway markers in in vitro and in vivo AD model with or without miR-4536-3p-I treatment. ( A ) Heat map shows the expression of PI3K/Akt pathway-related proteins and apoptotic markers in the in vitro AD model treated with or without miR-4536-3p-I treatment, ( B ) Network diagram illustrates the protein interactions within the PI3K/Akt signaling pathway, with line thickness representing the interaction strength between proteins. ( C - E ) Western blotting analysis of PI3K, p-Akt, and Akt expression in the cerebral cortex of WT or 5xFAD mice following miR-4536-3p-I treatment. ( F - H ) Western blotting analysis of PI3K, p-Akt, and Akt expression in the hippocampus of WT and 5xFAD mice following miR-4536-3p-I treatment. β-actin was used as the endogenous control. Results represent means ± SEMs ( n  = 4–6 per group). Statistical significance is indicated as follows: # p  < 0.05, ## p  < 0.01, and ### p  < 0.001 compared with the 5xFAD or 5xFAD + NC group Expression of PI3K/Akt/GSK3β signaling pathway markers in in vitro and in vivo AD model with or without miR-4536-3p-I treatment. ( A ) Heat map shows the expression of PI3K/Akt pathway-related proteins and apoptotic markers in the in vitro AD model treated with or without miR-4536-3p-I treatment, ( B ) Network diagram illustrates the protein interactions within the PI3K/Akt signaling pathway, with line thickness representing the interaction strength between proteins. ( C - E ) Western blotting analysis of PI3K, p-Akt, and Akt expression in the cerebral cortex of WT or 5xFAD mice following miR-4536-3p-I treatment. ( F - H ) Western blotting analysis of PI3K, p-Akt, and Akt expression in the hippocampus of WT and 5xFAD mice following miR-4536-3p-I treatment. β-actin was used as the endogenous control. Results represent means ± SEMs ( n  = 4–6 per group). Statistical significance is indicated as follows: # p  < 0.05, ## p  < 0.01, and ### p  < 0.001 compared with the 5xFAD or 5xFAD + NC group Fig. 10 Mechanisms of miR-4536-3p inhibition via the DBN1/PI3K/Akt/GSK3β signaling pathway. miR-4536-3p inhibition increases the DBN1 expression and activates the PI3K/Akt signaling pathway. DBN1 promotes neuronal protection, while p-Akt suppresses apoptosis and inhibits GSK3β activity. This cascade ultimately reduces tau hyperphosphorylation and alleviates AD symptoms. Therefore, miR-4536-3p-I may elicit a therapeutic effect on in vitro and in vivo AD models Mechanisms of miR-4536-3p inhibition via the DBN1/PI3K/Akt/GSK3β signaling pathway. miR-4536-3p inhibition increases the DBN1 expression and activates the PI3K/Akt signaling pathway. DBN1 promotes neuronal protection, while p-Akt suppresses apoptosis and inhibits GSK3β activity. This cascade ultimately reduces tau hyperphosphorylation and alleviates AD symptoms. Therefore, miR-4536-3p-I may elicit a therapeutic effect on in vitro and in vivo AD models

Alzheimer’s disease (AD), the main type of dementia, accounts for 60–80% of dementia cases [ 1 ]. It is a serious neurodegenerative disease with cognitive or behavioral impairments such as memory loss, difficulty with planning or problem solving, trouble in familiar task completion, and avoidance of work or social activities [ 2 ]. It is caused by various factors, including amyloid cascade, tau toxicity cascade, synaptic plasticity, and synapses loss [ 3 ]. It is also attributed to other risk factors, such as genetic factors, vascular diseases, increasing age, environmental factors, head injuries, and infections [ 4 ]. Amyloid pathogenesis, the major progression in AD, occurs with an altered cleavage of amyloid precursor protein (APP) to product amyloid-beta (Aβ) fibrils by β- and γ-secretase [ 5 ]. Aβ fibrils either directly induce the abnormal phosphorylation of tau protein or malfunctioning of neurons; consequently, these abnormalities trigger the formation of intracellular tau neurofibrillary tangles (NFTs) [ 6 ]. Subsequently, these occurrences activate the inflammatory response of microglia and induce neurotoxicity [ 5 , 6 ]. As demographic aging occurs and life expectancy rises, the worldwide incidence of AD is expected to increase substantially in developing countries, resulting in a costly disease burden [ 7 ]. Although clinical trials against AD are ongoing, more than 30 phase-3 trials have failed to show considerable cognitive improvements in patients with AD or occasionally caused harm even when amyloid plaques have been successfully reduced [ 8 – 11 ]. Nowadays, microRNAs (miRNAs) have emerged as novel agents for AD pathogenesis and potential drug candidates for AD treatment. miRNAs are small non-coding RNAs that modulate gene expression by binding to the 3ʹ-untranslated region (3ʹ-UTR) of the mRNA at the post-transcriptional level [ 12 ]. One nuclear miRNA gene transcribed by RNA polymerase II is cut into a precursor miRNA by Drosha and Di George critical region 8, and exportin-5 facilitates pre-miRNA transport into the cytoplasm [ 13 ]. In the cytoplasm, Dicer cleaves the pre-miRNA, generating a double-stranded mature miRNA [ 14 ]. The RNA-induced silencing complex incorporates single miRNA strands with Argonaute proteins, enabling this ribonucleoprotein complex to attach to target sequences within the 3′-UTR of mRNA [ 15 ]. miRNAs can influence numerous critical biological processes, including cell proliferation, cell growth, embryonic development, tissue differentiation, and apoptosis, by interacting with multiple target genes [ 16 ]. Some studies have reported that miRNAs, such as miR-200a-3p, miR-195, miR-338-5p, miR-34a-5p, miR-125b-5p, and miR-132, can be used as therapeutic reagents against AD by targeting β-secretase 1 that induced Aβ production [ 17 – 23 ]. In light of these findings, exploring additional miRNAs with potential regulatory roles in AD pathology has become increasingly important. Although the role of miR-4536-3p in neurodegenerative diseases remains unexplored, its dysregulation has been reported in various pathological conditions. For example, it is upregulated in the cerebrospinal fluid of patients with leptomeningeal metastasis [ 24 ]​, as well as in chemoresistant chronic myelocytic leukemia cells [ 25 ]​, while being downregulated following PM2.5 exposure in lung epithelial cells​ [ 26 ]. It has also been included in diagnostic miRNA panels for pancreatic cancer and endometriosis​ [ 27 , 28 ]. Notably, our preliminary microarray analysis revealed that miR-4536-3p is consistently upregulated in neural-induced human adipose-derived mesenchymal stem cells (NI-hADSCs) compared to undifferentiated hADSCs (Fig. S1 ), implying a potential role in neural differentiation. Supporting this, our recent study suggested that miR-4536-3p may regulate neuronal differentiation through the Wnt/MAPK signaling pathway in SH-SY5Y cells [ 29 ]. Furthermore, public small RNA-seq data from the Tissue Atlas Database indicate that miR-4536-3p is detectably expressed in human brain tissue (Fig. S2 C). Collectively, these findings highlight the biological relevance of miR-4536-3p in the central nervous system and support further investigation into its potential role in neurodegenerative disorders such as AD. Drebrin1 (DBN1), predominantly expressed in the brain, serves as a key intracellular modulator of neuronal morphogenesis [ 30 ]. It is involved in collateral axon branching in cortical layer II/III somatosensory neurons and essential for neuronal migration, neurogenesis, and synaptic plasticity [ 31 , 32 ]. Its reduced levels are recognized as a hallmark of several neurodegenerative disorders, including AD [ 32 ]. In an APP/PS1 AD mouse model, DBN1 overexpression improves cognitive function and mitigates neuropathological lesions [ 33 ]. Despite its critical role in AD pathology, its effect in the context of AD remains underexplored. In this study, we identified DBN1 as a target gene of miR-4536-3p. On the basis of this finding, we hypothesized that upregulating the DBN1 expression by using an miR-4536-3p inhibitor (miR-4536-3p-I) could offer therapeutic and neuroprotective benefits to an AD model. This research aimed to evaluate the effects of miR-4536-3p-I on in vitro and in vivo AD models by investigating its influence on apoptotic levels, neuronal marker expression, Aβ and tau pathology, cognitive function, and the phosphatidylinositol 3-kinase (PI3K)/Akt signaling pathway.

AD is a neurodegenerative disorder and the leading cause of dementia. It is characterized by cognitive decline, behavioral impairments, and pathological hallmarks such as Aβ plaques and tau neurofibrillary tangles. Although numerous clinical trials targeting amyloid plaques have been conducted, therapeutic advancements have been limited. In this study, we highlighted the potential of miR-4536-3p-I as a promising therapeutic strategy for AD. Our qPCR analysis, luciferase reporter assay, and western blotting analysis revealed that DBN1, a key regulator of neuronal morphogenesis and synaptic plasticity, was a direct target of miR-4536-3p. miR-4536-3p inhibition repressed apoptosis and protected against neuronal loss in in vitro and in vivo AD models. Moreover, the treatment with miR-4536-3p-I significantly reduced Aβ accumulation and tau hyperphosphorylation, which are the major pathological drivers of AD, in the brains of 5xFAD mice. MWM demonstrated that these changes were accompanied by cognitive improvements. Mechanistically, we confirmed that miR-4536-3p-I activated the PI3K/Akt/GSK3β signaling pathway, which helped inhibit apoptosis and alleviate AD symptoms (Fig.  10 ). Collectively, these findings suggested that miR-4536-3p inhibition shows therapeutic potential for AD by targeting the DBN1/PI3K/Akt/GSK3β signaling axis. The role of miRNAs in AD pathogenesis has been increasingly explored. miRNAs such as miR-20b-5p, miR-143-3p, and miR-148a-3p can regulate key AD-related pathways, including APP processing and tau phosphorylation, to improve neuronal function and cognitive outcomes [ 54 – 56 ]. Similarly, miR-204-3p, miR-135a-5p, miR-146a, and miR-181a can modulate cognitive impairment in AD [ 57 – 59 ]. However, miR-4536-3p has not been studied in this context. Our research is the first to demonstrate the therapeutic potential of miR-4536-3p-I in AD and to identify DBN1 as its target gene in this disease. DBN1 is critical for neuronal functions, including axon branching, synaptic plasticity, neuronal migration, and neurogenesis [ 31 , 32 ]. When upregulated, it promotes neuronal stem cell differentiation and porcine skeletal muscle satellite cell differentiation [ 60 , 61 ]. During brain development, it is predominantly expressed in the cortex, ventricular zone, and hippocampus [ 60 ]. Its reduced DBN1 levels are a hallmark of AD, and its overexpression improves cognitive function in AD mouse models [ 32 , 33 ]. Despite its recognized importance, the detailed mechanisms underlying its role in AD pathology remain poorly understood. In our study, qPCR analysis, luciferase assay, and western blotting analysis confirmed that DBN1 was directly targeted by miR-4536-3p in SH-SY5Y cells, HEK293T cells, and 5xFAD mouse brains. When miR-4536-3p was inhibited, DBN1 was upregulated, thereby reducing Aβ accumulation and tau phosphorylation in 5xFAD mice. Additionally, an increased DBN1 expression improved cognitive function and increased the levels of neuronal markers, including NeuN, NFH, MAP2, SYP, and Tuj1. Given that NeuN is a marker of mature neurons, the increased NeuN expression observed after miR-4536-3p inhibition is more likely indicative of neuronal preservation or differentiation rather than active neurogenesis. This suggests that miR-4536-3p inhibition supports neuronal integrity rather than directly promoting new neuron formation. Therefore, miR-4536-3p-I showed therapeutic potential for alleviating AD pathology by modulating the DBN1/PI3K/Akt/GSK3β signaling pathway. Despite these promising results, some discrepancies between mRNA and protein expression levels were observed. For instance, Tuj1 mRNA expression was most significantly increased at 1 pM miR-4536-3p-I, while its protein level was most notably elevated at 0.1 pM. This inconsistency may be attributed to post-transcriptional regulatory mechanisms, including mRNA stability, translational efficiency, and protein degradation, which are commonly reported in studies comparing transcriptome and proteome data [ 62 ]. Furthermore, the expression levels of several neuronal markers, including Tuj1, did not differ significantly between Aβ-treated and control SH-SY5Y cells. Although MTT assay indicated a ~ 30% reduction in cell viability after Aβ exposure, the remaining viable cells may have retained neuronal characteristics, resulting in unchanged marker expression. Moreover, SH-SY5Y cells are not fully differentiated neurons and may exhibit limited phenotypic plasticity in response to Aβ stress under the given treatment conditions [ 63 ]. This suggests that the absence of significant changes in neuronal marker expression could reflect the intrinsic limitations of the model rather than a true lack of biological effect. The PI3K/Akt signaling pathway plays a critical role in the brain by regulating neuronal survival, proliferation, apoptosis, and cell cycle progression [ 64 , 65 ]. As a key second messenger, PIP3 facilitates Akt activation through PDK1; as such, Akt can modulate downstream targets such as GSK3β, mTORC, FOXOs, and CREB [ 65 , 66 ]. In the CNS, an increased GSK-3β activity is directly associated with increased amyloid beta production and deposition, tau hyperphosphorylation, and neurofibrillary tangle formation; consequently, it contributes to AD progression [ 65 , 67 ]. Given this connection between the PI3K/Akt pathway and AD, several studies have explored its therapeutic potential by using specific compounds to target this pathway. For instance, Kai-Xin-San, Qingxin Kaiqiao Fang, and sulforaphane inhibit tau phosphorylation, neuronal apoptosis, and inflammation via PI3K/Akt/GSK3β signaling in in vitro and in vivo AD models [ 68 – 70 ]. Similarly, our study demonstrated that miR-4536-3p-I activated the PI3K/Akt/GSK3β signaling pathway in in vitro and in vivo AD models. The activation of this pathway reduced Aβ accumulation and tau phosphorylation, thereby improving the cognitive function of 5xFAD mice. Therefore, the PI3K/Akt signaling pathway is crucial in mitigating AD symptoms and is a potential therapeutic target. This study has several limitations. First, DBN1 expression did not differ significantly between WT and 5xFAD mice although studies have shown a decrease in DBN1 levels in AD models [ 33 ]. This discrepancy may be attributed to age-dependent variations in the expression of AD-related factors in the brain [ 33 , 71 , 72 ]. For instance, a previous study demonstrated that DBN1 expression was significantly reduced in the hippocampus of 12-month-old APP/PS1 mice, whereas no significant difference was observed in 4-month-old APP/PS1 mice compared to age-matched WT controls [ 33 ]. These findings suggest that the 4-month-old 5xFAD mice used in this study might be too young to exhibit significant changes in DBN1 expression. Nevertheless, the treatment with miR-4536-3p-I upregulated the DBN1 expression and improved cognitive function in 5xFAD mice, suggesting that DBN1 may act as a promoter of cognitive function. Second, tau phosphorylation was not significantly altered between WT and 5xFAD mice. This finding could be due to the nature of the 5xFAD model, which predominantly focuses on Aβ accumulation rather than tau pathology [ 73 ]. Thus, 5xFAD mice might not be an optimal model for studying tau phosphorylation. Third, while the MWM appropriately assessed spatial learning and memory, additional behavioral tests (e.g., novel object recognition, Y-maze) would have enabled a more comprehensive evaluation of cognitive and emotional function. Future studies should consider incorporating multiple behavioral paradigms to fully assess the therapeutic effects of miR-4536-3p inhibition.

In conclusion, we identified miR-4536-3p as a novel regulator in AD and showed the therapeutic potential of its inhibition. It directly targeted DBN1, a key neuronal modulator, and its inhibition reduced Aβ accumulation, tau hyperphosphorylation, and apoptosis. It also enhanced neuronal survival and cognitive function in in vitro and in vivo AD models. These effects were mediated by the activation of the PI3K/Akt/GSK3β signaling pathway. Thus, our findings demonstrated miR-4536-3p inhibition as a promising therapeutic strategy for AD and provided a basis for developing miRNA-based treatments targeting neurodegeneration.

Below is the link to the electronic supplementary material. Supplementary Material 1 Supplementary Material 1