Introduction

Alzheimer’s disease (AD), which is the most common type of dementia, accounts for 60–70% of dementia cases1. It is a neurodegenerative disease that involves the continuous impairment of cognitive and behavioral functions including attention, reasoning, memory, judgment, comprehension, and language2. It can be stimulated by several factors such as aging, female sex, low education level, frailty, family history, and presence of neurodegeneration markers2,3. Its fundamental pathology involves the accumulation of extracellular amyloid plaques because of the cleavage of β-amyloid (Aβ) peptides from amyloid precursor protein (APP) and the formation of intracellular Tau neurofibrillary tangles (NFTs) via tau hyperphosphorylation1,2,4. Although it is a serious disease, no definite cure is available except for symptomatic treatments such as cholinesterase inhibitors, N-methyl-D-aspartate receptor antagonist, and lecanemab, which can suppress dementia symptoms for a short period or induce some adverse events1,2,5. Clinical trials of new drugs for AD target anti-amyloid, anti-neuroinflammation, anti-tau, neuroprotection, cognitive enhancement, or behavioral psychological symptoms of dementia1. MicroRNAs (miRNAs), considered as novel agents of intercellular communication, offer new perspectives on previously identified AD therapeutic strategies6,7.

miRNAs are small RNA molecules that have a length of approximately 22 nucleotides and regulate gene expression8. They mainly function by binding to the 3ʹ-untranslated region (UTR) of messenger RNA (mRNA), thereby repressing protein translation or degrading mRNA molecules9. Within the cell nucleus, miRNAs are initially transcribed as primary transcripts (pri-miRNA) by RNA polymerase II10. Then, pri-miRNAs are processed by Drosha to form hairpin structures called precursor miRNAs (pre-miRNAs), which are then exported to the cytoplasm by exportin 511. In the cytoplasm, they are cleaved by another enzyme called Dicer; consequently, this process produces double-stranded miRNA molecules, which are subsequently incorporated into Argonaute (Ago) proteins10,12. Once loaded onto Ago proteins, one strand of a mature miRNA forms an RNA-induced silencing complex, which serves as a guide to target specific mRNA molecules for silencing13.

miRNAs can target hundreds of mRNAs because of the imperfect complementarity needed for binding14. As such, they play essential roles in various biological processes, including apoptosis, cell cycle regulation, proliferation, and differentiation15,16. For instance, miR-484 and miR-96 induce apoptosis by targeting SESN2 and XIAP, respectively17,18. miR-34a-5p and miR-200a regulate apoptosis19,20. Additionally, several miRNAs modulate AD pathogenesis by regulating AD-related genes. For example, miR-298 and miR-346 can target the human APP21,22. miR-33 influences Aβ metabolism, a primary cause of AD23. However, specific miRNA-targeted mRNAs and their roles in apoptosis and neuroprotection in AD remain unclear.

In our previous study, neural-induced human adipose tissue-derived stem cells (NI-hADSC) exhibit functional characteristics similar to neuronal cells when they are exposed to forskolin and basic fibroblast growth factor24,25. Although stem cell-derived exosomes containing miRNAs have been explored as a potential treatment for AD, studies have yet to investigate the use of NI-hADSC-derived miRNAs for this purpose26,27,28,29. Following a microarray analysis, miR-937-3p could be a candidate and its expression was significantly increased in Aβ-treated SH-SY5Y cells. Thus, we hypothesized that miR-937-3p could influence neuroprotection and AD pathology25. Although miR-937-3p has been studied in cancer, its role in neurodegenerative disease or neuronal cell death remains unclear30,31,32. A database search (PubMed, miRBase v22.1, HMDD v4.0, as of May 9, 2025) revealed no reported associations with AD or neuronal apoptosis, highlighting the novelty of this study. Our findings revealed that miR-937-3p inhibition exerted neuroprotective effects on the Aβ-induced AD model by upregulating Netrin1 (NTN1), a gene involved in axon guidance, and by suppressing Aβ-induced apoptosis. Furthermore, we revealed that miR-937-3p inhibition elicited a neuroprotective effect on the in vitro AD model. Therefore, miR-937-3p inhibition plays a therapeutic role in AD by promoting the NTN1 expression and inhibiting the apoptosis pathway.

Materials and methods

Compounds and experimental design

Aβ1−42 (Cat. No. AS-20276, ANASPEC, Fremont, CA, USA) was first dissolved in 1% NH4OH (Cat. No. A-6899, Sigma-Aldrich, St. Louis, MO, USA) to prepare a 500 µM stock solution. This stock solution was then diluted in a medium to obtain a final concentration of 500 nM Aβ1−42. In all experiments, the final NH4OH concentration was < 0.001%.

The miR-937-3p inhibitor (miR-937-3p-I, UAGGCGCGAGACUGAGAGACGG; Bioneer, Daejeon, Republic of Korea), the miR-937 mimic (miR-937-3p-M, AUCCGCGCUCUGACUCUCUGCC; Bioneer), the negative control inhibitor (NC-I; Bioneer), and the NC mimic (NC-M; Bioneer) were dissolved in ultra-pure distilled water (Invitrogen Co., Waltham, MA, USA) until their concentrations reached 0.1, 0.5, and 1 nM, respectively.

The human neuroblastoma cell line SH-SY5Y (RRID: CVCL_0019l ATCC® CRL-2266) and the human embryonic kidney cell line HEK293T (ATCC® CRL-11268) were purchased from ATCC (Manassas, VA, USA) and cultured in Dulbecco’s modified Eagle’s medium (Cat. No. SH30243.01, Cytiva, Marlborough, MA, USA) supplemented with 10% fetal bovine serum (FBS; Cat. No. SV3020702, Cytiva), 0.2% amphotericin B (Cat. No. 15290-018, Thermo Fisher Scientific Co., Waltham, MA, USA) and 1% penicillin-streptomycin (Cat. No. LS202-02, WELGENE, Gyeongsan-si, Republic of Korea). The cells were maintained at 37℃ in a humidified atmosphere containing 5% CO2 and 95% air33. We confirmed that the cell lines used in our study were commercially purchased from ATCC, which provides authenticated cell lines for research use.

To establish an in vitro AD model, we selected SH-SY5Y cells, which a widely used neuronal cell line for AD research, with 500 nM Aβ1−42. The concentration of Aβ was determined based on MTT assay and previous research that used Aβ concentrations below 10 µM in SH-SY5Y cells34,35,36,37. After 1 h, the miR-937-3p-I was added at a concentration of 0.1, 0.5, or 1 pM. The cells were incubated with or without Aβ1−42 and the miR-937-3p-I for 24 h before the subsequent experiments.

Bioinformatic analysis

Genes with complementary sequences to miR-937-3p were identified using TargetScan (https://www.targetscan.org/vert_80/) to determine miR-937-3p target genes. The target genes with the highest degree of complementarity to miR-937-3p were selected, and their inhibitor was treated in SH-SY5Y cells. Gene ontology (GO) analysis was conducted to characterize the biological function of miR-937-3p and explore the functional role of miR-937-3p target genes in terms of biological processes. Differentially expressed genes were subjected to GO enrichment analysis in the cluster Profiler R package, and gene length bias was corrected. In this analysis, a biological system defined by Kyoto Encyclopedia of Genes and Genomes (KEGG) was used with miRWalk (http://mirwalk.umm.uni-heidelberg.de/)38,39.

Antibody array

Antibody array analysis was conducted to validate the effects of the miR-937-3p-I on specific proteins. The cells were collected following treatment with Aβ and the miR-937-3p-I. FullMoonBio cell signaling phosphor arrays were performed by E-BIOGEN Inc. (Seoul, Republic of Korea), comparing the Aβ-treated group to the group treated with the miR-937-3p-I after Aβ treatment. The results were analyzed using ExDEGA, MeV, DAVID (https://david.ncifcrf.gov/), and STRING (https://string-db.org/) software.

Cloning and luciferase assay

The pmirGLO dual luciferase miRNA target expression vector (Cat. No. E1330, Promega, MA, USA) was used to confirm the binding activity of miR-937-3p to the 3ʹ-UTR of NTN1. The 3ʹ-UTR of NTN1, containing the miR-937-3p target site, was amplified and inserted into the pmirGLO vector (pmirGLO-NTN1). A mutant version of the miR-937-3p target site within the pmirGLO-NTN1 vector was generated using an EZchange™ site-directed mutagenesis kit (Cat. No. EZ004S, Enzynomics, Daejeon, Republic of Korea). The relative luciferase activity was assessed using a dual-luciferase reporter assay system (Cat. No. E2940, Promega), in accordance with the manufacturer’s instructions40. HEK293T cells were co-transfected with either wild-type or mutant pmirGLO-NTN1 plasmids and either the miR-937-3p-M or the NC by Lipofectamine 3000 (Cat. No. 2412049, Invitrogen Co.). After 48 h, they were harvested, and luciferase activity was measured. Firefly luciferase luminescence was detected after passive lysis buffer was added. The relative activity of expression was determined by calculating the ratio of firefly to Renilla luciferase signals41. The vector map used in this study was designed with the VectorBee software (Fig. S1).

RNA isolation and quantitative PCR (qPCR)

After Aβ and miR-937-3p-I treatment, total RNA was extracted from SH-SY5Y cells by using TRIzol (Cat. No. 9109, Takara Bio Co., Shimogyo-ku, Japan) as previously described31. cDNA was synthesized from the extracted RNA with a PCR instrument (Takara Bio Co.)32.

For microRNA qPCR analysis, cDNA was synthesized using the miRCURY LNA RT Kit (Cat. No. 339340, QIAGEN, Venlo, Netherlands), and microRNA qPCR was performed using the miRCURY LNA SYBR Green PCR Kit (Cat. No. 339345, QIAGEN) according to the manufacturer’s protocol. U6 was used as the endogenous reference control for normalization.

The upregulation of the target genes in the SH-SY5Y cells treated with the miR-937-3p-I was assessed via qPCR. SYBR Green Premix Ex Taq (Cat. No. RR420, Takara Bio Co.) and LightCycler 480 II (Roche Holding AG., Basel, Swiss) were used to amplify NTN1 or neuronal markers such as MAP2, Tuj1, NEFL, or NEFM at an annealing temperature of 60 °C. GAPDH was used as the endogenous reference control for normalization. Each sample was analyzed in triplicate. All primers listed in Table 1 were purchased from CosmoGenetech (Seoul, Republic of Korea), Macrogen (Seoul, Republic of Korea), and Bioneer.

Table 1 Sequence of qPCR primers.
Full size table

Fluorescence-activated cell sorting (FACS) analysis

FACS analysis was performed in accordance with the manufacturer’s instructions to examine apoptosis regulation in cells44. In brief, the cells were washed with cold phosphate-buffered saline (PBS, Biosesang, Yongin, Republic of Korea) and resuspended in 1X binding buffer. They were incubated with FITC Annexin V and PI (Cat. No. 556547, BD Biosciences, Franklin Lakes, NJ, USA) in the dark for 15 min. They were examined using FACSCanto II (BD Biosciences), and results were processed using BD FACSDiva software (BD Biosciences). All experiments were analyzed at least thrice.

Western blot analysis

Western blot analysis was performed to investigate the signaling pathway with Aβ or miR-937-3p-I in accordance with previously described methods43,45,46. In brief, the cells (3 × 105 cells/mL) were cultured with Aβ1−42 and/or miR-937-3p-I, and proteins were isolated using a lysis buffer (0.5 M EDTA, 10% Nonidet P-40, 50 mg/mL leupeptin, 0.1 M Na3VO4, 1 M NaF, 50 mg/mL aprotinin, 1 M Tris pH 7.5, 1 M NaCl, 100% glycerol, and 0.2 M PMSF). Protein concentrations were measured using a BCA protein assay kit (Cat. No. 23227, Thermo Fisher Scientific Co.) in accordance with the manufacturer’s instructions47. Equal amounts of protein (15 µg/10 µL) from each group were separated via 10% or 6% sodium dodecyl sulfate-polyacrylamide gel electrophoresis and transferred to a polyvinylidene difluoride membrane (Cat. No. IPVH00010, Merck Millipore Co., Burlington, MA, USA). All the primary and secondary antibodies are listed in Table 2. Results were visualized using an Immobilon Crescendo Western HRP substrate (Cat. No. WBLUR0500, EMD Millipore Co.) and quantitatively analyzed with ImageJ software. All experiments were conducted at least in triplicate.

Table 2 Antibody list.
Full size table

Immunocytochemistry

Cells were seeded onto poly-L-lysine-coated aclar plastic coverslips at a density of 3 × 10⁵ cells/mL. Following treatment of miR-937-3p-I (1 pM) after Aβ exposure (500 nM), the SH-SY5Y cells were fixed with 4% paraformaldehyde (Cat. No. SM-P01-100, GeneAll, Seoul, Republic of Korea) for 15 min. Permeabilization was then performed using 0.5% Triton X-100 (Cat. No. X100, Sigma-Aldrich) for 20 min, followed by blocking with 10% normal goat serum (Cat. No. S-1000, Vector Laboratories, Inc., Newark, CA, USA) for 30 min, as previously described43,45.

For immunostaining, cells were incubated with primary antibodies targeting microtubule-associated protein 2 (MAP2, 1:200; Cat. No. 8707, Cell Signaling Technology) for 1.5 h. Alexa 488-conjugated goat anti-rabbit secondary antibody (Cat. No. A1108, Invitrogen Co.) was then applied. Nuclei were counterstained with 4′,6-diamidino-2-phenylindole (1 µg/mL; Cat. No. D1306, Life Technologies Corporation, CA, USA). Imaging was conducted using an Axio Vert. A1 microscope (Carl Zeiss, Germany).

Statistical analysis

Data were statistically analyzed using GraphPad Prism® 5.0 (GraphPad Software Inc., San Diego, CA, USA). The data are presented as means ± standard error of the mean (SEM). A two-tailed unpaired Student t-test was used to assess the significance of differences between two groups, such as NC versus miR-937-3p. For comparisons among three or more groups, such as the Aβ group versus the Aβ + miR-937-3-I group, one-way ANOVA was applied, followed by the Tukey’s post-hoc test. For multiple comparisons, two-way ANOVA with the Bonferroni post-hoc test was performed. Data with p < 0.05 were considered statistically significant, and the significance level was indicated as follows: *p < 0.05, **p < 0.01, and ***p < 0.001 compared with the control group; #p < 0.05, ##p < 0.01, and ###p < 0.001 compared with the Aβ-treated group.

Results

Identification of miR-937-3p in the neuroprotection of AD dataset

A microarray analysis was performed to compare the expression levels of miRNA candidates between hADSC and NI-hADSC prepared from our previous studies24. The miRNAs were selected when their expression was at least 1.3 times higher in NI-hADSC than in hADSC across three samples. In addition, we considered that the miRNA candidates were not previously studied in the context of apoptosis or neuronal protection. Among them, miR-937-3p was singled out and further examined to evaluate its anti-apoptotic and neuroprotective effects against Aβ1−42 toxicity. Heatmap and scatter plots indicated that the miR-937-3p expression was consistently higher in NI-hADSC than in hADSC in the three samples (Fig. S2).

We conducted a biological analysis of miR-937-3p by using miRWalk to explore the related pathways, such as chemical information, biological functions, and molecular functions. KEGG pathway analysis revealed that miR-937-3p was significantly related to neuronal differentiation and neuroprotective pathways, including axon guidance, Wnt signaling, and dopaminergic synapse pathways (Fig. S3A). GO analysis indicated that miR-937-3p participated in cell proliferation and neuronal differentiation, which were related to neuron projection, nervous system development, cell number homeostasis, and axon extension (Fig. S3B-D). These data suggest that miR-937-3p has the potential for neuroprotective properties.

Next, we conducted microRNA qPCR to evaluate miR-937-3p expression levels among control group, Aβ-treated group, and Aβ + miR-937-3p-I-treated group. The expression of miR-937-3p was significantly increased in the Aβ-treated group compared to the control group. However, miR-937-3p expression was notably reduced following miR-937-3p-I treatment in Aβ-treated SH-SY5Y cells (Fig. 1A). These findings suggest that elevated miR-937-3p levels are associated with Aβ-induced cell toxicity, and that reducing its expression may have a therapeutic effect in Aβ-treated SH-SY5Y cells.

To explore this mechanism, we performed an antibody array to compare differences in protein expression between the Aβ-treated group and the miR-937-3p-I-treated group after Aβ treatment in the cells. The heatmap showed that specific proteins were upregulated or downregulated in the Aβ with miR-937-3p-I treatment compared with the Aβ group (Fig. 1B). Following the results, we identified the top 15 disease-related proteins using DAVID and EXDEGA software. The proteins significantly upregulated in the Aβ + miR-937-3p-I group compared with the Aβ group were implicated in AD progression; furthermore, the AD-associated proteins were identified from the analysis (Fig. 1C-D). Additionally, KEGG pathway and GO analysis showed differential protein expression between the Aβ group and the Aβ with miR-937-3p-I treatment group (Fig. 1E-H), with some proteins potentially negatively affecting the apoptotic process (Fig. 1F). Taken together, the miR-937-3p-I treatment could affect the expression of AD-related and apoptotic protein in the in vitro AD models.

Fig. 1
figure 1

Comparative biological screening of the Aβ group and the Aβ with miR-937-3p-I group using an antibody array and gene expression profiling. (A) The expression level of miR-937-3p was elevated in the Aβ-treated group but significantly reduced following miR-937-3p-I treatment in the Aβ-treated group (n = 5 per group). Results represent means ± SEMs. Statistical significance is indicated as follows: ***p < 0.001 compared with the control; ###p < 0.001 compared with the Aβ. (B) Heat map illustrates the differential protein expression between the Aβ group and the Aβ with miR-937-3p-I group. Red color indicates a value of fold change more than 1.1. (C) The distribution graph depicts the enriched disease processes associated with differentially expressed proteins. Red bars represent the number of upregulated proteins linked to specific diseases, while blue bars represent the number of downregulated proteins. The black dot-line indicates the statistical significance of these enrichments as –Log10(p-value). (D) A network diagram shows the interactions among proteins specifically related to AD. KEGG pathway (E) and GO (F–H) analyses were performed. The KEGG pathways present differentially expressed proteins identified in the heat map. GO analyses categorize the differentially expressed proteins such as biological processes, cellular components, and molecular functions.

Full size image

miR-937-3p targets NTN1 and modulates its expression

We identified the potential target genes with complementary base sequences to miR-937-3p via the TargetScan website and visualized the candidate target genes by using Cytoscape (Table 3 and Fig. 2A). NTN1, which participates in axon guidance, was selected for further investigation to assess upregulation after miR-937-3p-I treatment.

qPCR analysis revealed that the mRNA expression of NTN1 was significantly increased with miR-937-3p-I treatment (1 pM) compared with NC-I. In contrast, miR-937-3p-M (1 pM) significantly reduced the mRNA expression of NTN1 compared to the NC-M in SH-SY5Y cells (Fig. 2B). Similarly, western blott analysis suggested that inhibition of miR-937-3p (25 nM) upregulated NTN1 protein expression, while miR-937-3p-M (25 nM) downregulated NTN1 protein levels compared to their respective NC in SH-SY5Y cells (Fig. 2C-F, S4).

To further assess the role of NTN1 in AD, we analyzed its expression in Aβ-treated SH-SY5Y cells following miR-937-3p-I treatment. The results indicated that mRNA expression of NTN1 was also upregulated in Aβ-treated cells with miR-937-3p-I treatment compared to cells treated with Aβ alone (Fig. 2G).

To investigate whether miR-937-3p directly binds to the 3′-UTR of NTN1 to regulate its expression, we performed a luciferase reporter assay. Figure 2H shows the complementary and mutant regions of NTN1 relative to miR-937-3p. The relative luciferase activity significantly decreased co-transfected with the miR-937-3p-M (25 nM) and pmiGlo-NTN1-WT in HEK293T cells, indicating direct binding of miR-937-3p to the 3′-UTR of NTN1. In contrast, the luciferase activity remained unchanged in cells co-transfected with the miR-937-3p-M (25 nM) and pmiGlo-NTN1-Mut, which contained mutations in the miR-937-3p binding site (Fig. 2I). These findings suggest that miR-937-3p can bind to the 3′-UTR of NTN1, thereby repressing its mRNA translation or promoting degradation. Therefore, inhibition of miR-937-3p mediates the upregulation of NTN1 expression by alleviating this post-transcriptional repression.

Fig. 2
figure 2

Target gene analysis of miR-937-3p. (A) Identification of miR-937-3p target genes using TargetScan and Cytoscape software. (B) mRNA expression levels of NTN1 in SH-SY5Y cells treated with miR-937-3p-I, miR-937-3p-M, or NC, assessed by qPCR (n = 6 per group). GAPDH served as an endogenous control. (C–F) Protein expression levels of NTN1 in SH-SY5Y cells treated with miR-937-3p-I, miR-937-3p-M, or NC, analyzed by western blot (n = 5 per group). (G) mRNA expression levels of NTN1 following treatment with Aβ alone or in combination with miR-937-3p-I in SH-SY5Y cells, assessed by qPCR (n = 6 per group). (H) The predicted binding site of miR-937-3p on the 3ʹ-UTR of NTN1 mRNA. Complementary sequences are highlighted in red. (I) Relative luciferase activity in cells co-transfected with either NC-M or miR-937-3p-M and the pmiGlo-NTN1 reporter construct. A significant decrease in activity was observed with the wild-type reporter, but not with the mutant reporter. All experiments were independently repeated, and technical replicates were also performed to ensure the reliability of the results. Results represent means ± SEMs. Statistical significance is indicated as follows: *p < 0.05; **p < 0.01; and ***p < 0.001 compared with the NC; ###p < 0.001 compared with the Aβ.

Full size image
Table 3 Gene expression profiling of miR-937-3p.
Full size table

miR-937-3p inhibitor represses Aβ-induced apoptosis of SH-SY5Y cells

To establish an in vitro AD model, we investigated the viability of SH-SY5Y cells after Aβ treatment in a dose-dependent manner by using the MTT assay (Fig. S5A). On the basis of MTT assay results, we selected a final concentration of 500 nM Aβ for subsequent experiments. To verify the aggregation status and morphological characteristics of Aβ1–42, we performed dynamic light scattering (DLS) and transmission electron microscopy (TEM) analyses. The DLS histogram, derived from regularized autocorrelation data, revealed the presence of Aβ fibrils (166.2 ± 16.9 nm) and larger Aβ aggregates (1122 ± 165.9 nm) in the prepared samples (Fig. S5B). TEM imaging further confirmed the coexistence of Aβ oligomers, fibrils, and amorphous aggregates (Fig. S5C). The observed fibrillar structures were consistent with previous reports of Aβ1–42 morphology visualized by TEM48,49,50,51. Collectively, these results provide direct evidence supporting the AD-like pathological features of the in vitro model used in this study.

The repression of apoptosis by the miR-937-3p-I in the in vitro AD model was explored via flow cytometry. The results showed that early apoptosis increased in the Aβ-treated group compared with that in the control group. However, the percentage of early apoptosis in the miR-937-3p-I-treated groups after Aβ treatment significantly decreased compared with that in the Aβ-treated group regardless of the miR-937-3p-I concentration. Similarly, the total apoptosis (including early and late apoptosis) in the miR-937-3p-I-treated groups was significantly reduced compared with that in the Aβ-treated group regardless of inhibitor concentrations (Fig. 3A-B). Therefore, the miR-937-3p-I might inhibit Aβ-induced apoptosis.

Fig. 3
figure 3

The detection of apoptosis using flow cytometry. The apoptosis levels in the cells were evaluated using flow cytometry (A) and calculated as a percentage of apoptosis (B). The percentage of early apoptosis and total apoptosis induced by Aβ was significantly reduced in the miR-937-3p-I-treated group compared with the Aβ-treated group (n = 5 per group). Results represent means ± SEMs. All experiments were independently repeated, and technical replicates were also performed to ensure the reliability of the results. Statistical significance is indicated as follows: **p < 0.01 and ***p < 0.001 compared with the control; ##p < 0.01 and ###p < 0.001 compared with the Aβ.

Full size image

We performed western blot analysis to investigate whether the miR-937-3p-I reduced Aβ-induced apoptosis. Figure 4A-B showed that the miR-937-3p-I (1 pM) upregulated the protein expression of Mcl-1, an anti-apoptotic marker, compared with the Aβ-treated group. Although the protein expression of Bcl-2, another anti-apoptotic marker, increased with miR-937-3p-I treatment, this finding was not significant. Interestingly, the Bax/Bcl-2 ratio was significantly reduced at 0.5 pM miR-937-3p-I treatment under Aβ-induced cell death, which indicated a shift toward cell survival. Additionally, the protein expression levels of pro-caspase-9, pro-caspase-7, pro-caspase-3, and pro-PARP, which are downregulated when their cleaved forms are upregulated, increased at certain concentrations of the miR-937-3p-I compared with those in the Aβ-treated group (Fig. 4C-D). Conversely, cleaved-caspase-3 expression was reduced following miR-937-3p-I treatment, and the cleaved-caspase-3/pro-caspase-3 ratio was significantly decreased, particularly at 0.5 pM miR-937-3p-I treatment. All un-cropped gel images were shown in the supplementary data (Fig. S6).

To further investigate the functional role of miR-937-3p, we examined the effects of miR-937-3p overexpression (miR-937-3p-M) in Aβ-treated SH-SY5Y cells. The expression of NTN1, a target gene of miR-937-3p, was significantly reduced in the Aβ + miR-937-3p-M group compared with the control group (Fig. S7). Additionally, Bax expression, a key pro-apoptotic marker, was markedly elevated in the Aβ + miR-937-3p-M group compared with both the control and Aβ-only groups. These findings further support the hypothesis that miR-937-3p regulates NTN1 expression and mediates apoptosis in Aβ-treated cells.

To determine whether NTN1 mediates the anti-apoptotic effects observed with miR-937-3p-I treatment, we performed NTN1 knockdown experiments in SH-SY5Y cells treated with Aβ and miR-937-3p-I. qPCR analysis confirmed that si-NTN1 significantly reduced NTN1 mRNA levels (Fig. S8A). Consistently, western blot analysis showed a notable reduction in NTN1 protein expression following si-NTN1 treatment in the presence of Aβ and miR-937-3p-I (Fig. S8B-C). Furthermore, NTN1 knockdown significantly decreased Mcl-1 expression, an anti-apoptotic marker, under these conditions (Fig. S8D-E). These results demonstrate that the anti-apoptotic effects induced by miR-937-3p-I in Aβ-treated cells are modulated by NTN1 expression. All un-cropped gel images were shown in the supplementary data (Fig. S9).

Fig. 4
figure 4

The regulation of miR-937-3p-I in the apoptotic cell death. Changes in apoptotic proteins after treatment of Aβ or/and miR-937-3p-I were assessed through western blot analysis (A,C) and the protein expressions were quantified (B,D). The major regulators of mitochondrial pathway, such as Bcl-2, Mcl-1, and Bax, and downstream effectors of apoptosis, including pro-caspase-9, 7, 3, and cleaved-caspase-3 as well as pro-PARP, were analyzed (n = 6 per group). The protein expression was normalized to the control group. β-actin was used as the control. Results represent means ± SEMs. All experiments were independently repeated, and technical replicates were also performed to ensure the reliability of the results. Statistical significance is indicated as follows: *p < 0.05, **p < 0.01, and ***p < 0.001 compared with the control; #p < 0.05, ##p < 0.01, and ###p < 0.001 compared with the Aβ.

Full size image

miR-937-3p inhibitor protects neurons against Aβ toxicity in SH-SY5Y cells

Western blot analysis, qPCR, and ICC were performed to confirm the expression of neuronal markers. The results showed that the protein expression levels of the neuronal markers, NeuN, NFH, Tuj1, and SYP, were upregulated with miR-937-3p-I treatment following Aβ exposure (Fig. 5A-B, Fig. S10). Additionally, the expression of MAP2, another neuronal marker, was promoted by miR-937-3p-I treatment regardless of concentration compared with that in the Aβ-treated group. The expression levels of Tuj1, NEFL, and NEFM, which are also neuronal markers, significantly increased at certain concentrations of miR-937-3p-I treatment after Aβ exposure in SH-SY5Y cells (Fig. 5C-F). Furthermore, ICC results indicated that MAP2 expression was reduced in Aβ-treated cells compared with the control; however, it was significantly increased with miR-937-3p-I (1 pM) treatment following Aβ exposure (Fig. 5G-H). Nevertheless, neurite length did not change in response to either Aβ or miR-937-3p-I treatment (Fig. 5I). Therefore, miR-937-3p-I elicited a neuroprotective effect on the Aβ-induced AD model.

Fig. 5
figure 5

Neuroprotective effects of miR-937-3p-I in Aβ-induced cell death. The protein expression of neuronal markers (NeuN, NFH, Tuj1, and SYP) was investigated by western blot analysis (A) and calculated the levels (B) (n = 6 per group). The level of genes, which are neuronal markers, was investigated via qPCR (C–F) (n = 5 per group). The expression of neuronal markers increased with the miR-937-3p-I treatment in Aβ-induced cells compared with that in the Aβ-treated group. (G) ICC showed MAP2 (green) expression treated with miR-937-3p-I after Aβ exposure (n = 5 per group). (H) MAP2-positive cells were quantified as relative to total nuclei (DAPI, blue). (I) Neurite length calculated and remained unchanged following Aβ or miR-937-3p-I treatment. All experiments were independently repeated, and technical replicates were also performed to ensure the reliability of the results. Results represent means ± SEMs. Statistical significance is indicated as follows: *p < 0.05, **p < 0.01, and ***p < 0.001 compared with the control; #p < 0.05, ##p < 0.01, and ###p < 0.001 compared with the Aβ.

Full size image
Fig. 6
figure 6

The mechanisms of miR-937-3p-I via NTN1 regulation in AD. miR-937-3p-loaded RISC mediates gene silencing via mRNA cleavage and degradation or translational repression, depending on the complementarity between the miRNA and the target mRNA transcript, such as NTN1. Following our results, the expression of NTN1 could downregulate the cleaved caspase-9, -3, -6, and − 7, and PARP, while upregulating the levels of pro-caspase-9, -3, -6, and − 7, and pro-PARP, finally reducing apoptosis in the cells. In addition, NTN1 could regulate anti-apoptotic proteins, such as Bcl-2 and Mcl-1, as well as the pro-apoptotic protein Bax. These findings suggest that miR-937-3p-Is may have a therapeutic effect in the in vitro AD model.

Full size image

Discussion

In AD, numerous genes are affected, leading to neuronal loss and apoptosis as the disease progresses52. They cause memory impairment and gene dysregulation, causing AD-related changes in the brain53. While miRNAs can bind and regulate multiple target genes, attempts to use them for modulating neuronal genes in AD have been poorly executed. In this study, we showed that miR-937-3p inhibition can prevent the miR-937-3p-mediated mRNA degradation or translational repression of NTN1. In addition, miR-937-3p inhibition could upregulate anti-apoptotic factors (Bcl-2 and Mcl-1), pro-caspase, and pro-PARP (Fig. 6). Furthermore, miR-937-3p inhibition likely promoted the neuroprotective effects on the in vitro AD model by increasing the expression of neuronal markers (NeuN, Tuj1, NFH, SYP, MAP2, NEFL, and NEFH).

The relationship between miRNAs and apoptosis in several diseases has been studied. For instance, miR-211 inhibits apoptosis in a cerebral ischemia injury model54. miR-34a-5p is a key modulator of apoptosis in myocardial infarction, and miR-203a-3p targets MYD88 and consequently reduces apoptosis in osteoarthritis55,56. However, few studies have investigated the role of miRNAs in apoptosis in AD15,57,58. Our study provided new insights and demonstrated that miR-937-3p-I could modulate apoptosis, potentially aggravating AD pathogenesis, as shown by FACS and western blot analysis.

miRNAs can target multiple genes participating in biological processes, such as cell proliferation, differentiation, and apoptosis. Because of these characteristics, certain miRNAs may serve as therapeutic targets for AD. For instance, miR-129-5p and miR-20b-5p have been identified as potential biomarkers in AD pathogenesis9,59. miR-433 regulates JAK2, and miR-193a-3p targets PTEN in AD pathogenesis60,61. Similarly, miR-650 and miR-29a can modulate AD pathogenesis by binding to cyclin-dependent kinase 5 or regulating Aβ deposition62,63. Moreover, miR-455-3p improves cognitive and synaptic functions in AD64. Our research indicated that miR-937-3p could be inhibited and used as a therapeutic target for AD by reducing apoptosis and neuronal loss.

Several studies have suggested the role of miR-937-3p in cancer and neurologic disorders30,31,32. For example, miR-937 promotes lung cancer cell proliferation32,65 and modulates apoptosis and proliferation in breast cancer by targeting specific genes30,31. miR-937 also exhibits anti-inflammatory effects by targeting IL-1B66. However, limited studies have explored the effects of miR-937-3p on AD67. The present study is the first to suggest that miR-937-3p might target NTN1, a neuronal gene, and regulate apoptosis and neuronal protection.

NTN1 is part of the netrin family of proteins, which are involved in neuronal axon guidance by modulating microtubule function68,69. NTN1 is also a therapeutic target for AD because it disrupts Aβ and improves cognitive impairment in PDAβPPSwe/Ind mice70. It has also been studied as a target gene of miR-let-7 in dorsal root ganglions71. Our research confirmed that NTN1 is a target gene of miR-937-3p via qPCR analysis and luciferase assay42,69. Given that NTN1 has been reported to inhibit an initiation of apoptosis42,69, we further investigated whether its upregulation via miR-937-3p-I could exert an anti-apoptotic effect in an in vitro AD model. FACS analysis showed that Aβ-induced early and total apoptosis was reduced by the miR-937-3p-I regardless of its concentration. Additionally, the expression of Mcl-1, an anti-apoptotic marker, increased with the miR-937-3p-I in the in vitro AD model. Similarly, the expression levels of pro-caspase 7, pro-caspase 3, and pro-PARP, which are typically downregulated when cleaved forms are upregulated, increased with the miR-937-3p-I in the in vitro AD model. Therefore, as demonstrated in other studies, NTN1 upregulated by miR-937-3p inhibition could prevent Aβ-induced apoptosis.

In this study, we observed that miR-937-3p expression was significantly increased in the Aβ-treated group, and NTN1 was identified as a direct target of miR-937-3p. However, the mRNA expression of NTN1 did not show a significant decrease in Aβ-treated cells compared to the control group (Fig. 2G). Given that previous studies suggested NTN1 functions upstream of Aβ72,73, this result indicated that NTN1 expression might not be directly regulated by Aβ. Instead, our findings implied that NTN1 expression is primarily modulated by miR-937-3p activity, rather than Aβ itself.

miRNAs are controversial molecules because they can target multiple genes in biological processes. For instance, miR-331-3p and its inhibitor elicit therapeutic and neuroprotective effects through different mechanisms in AD74,75. Similarly, both miR-200c and its inhibitor could serve as therapeutic targets, given that miR-200c expression is decreased in 5xFAD mice but upregulated in APP/PS1 mice53. These findings suggested that miRNAs might modulate multiple target sites. In the present study, miR-937-3p inhibition showed therapeutic and neuroprotective effects on the AD model by upregulating the NTN1 expression, so that it could be a beneficial and therapeutic agent in AD. However, given the dual roles of miRNAs in AD pathology, it remains to be determined whether miR-937-3p may exert different effects in distinct disease stages or cellular environments. Further investigation is required to elucidate whether miR-937-3p inhibition could have long-term neuroprotective effects in in vivo AD models or whether it may modulate other signaling pathways associated with disease progression.

The present study had several limitations. Firstly, we did not assess cognitive function in an in vivo AD model using miR-937-3p-I. As the 3ʹ-UTR of human NTN1 contains complementary sequences for miR-937-3p, while the mouse and rat NTN1 3ʹ-UTRs do not, we limited our investigation of miR-937-3p-I’s therapeutic effects to an in vitro AD model. Our next step is to evaluate the effects of miR-937-3p-I in AD mouse models and to explore whether it can target other genes to mediate therapeutic effects in vivo. Secondly, we did not directly investigate the regulatory mechanism between NTN1 and Aβ. While we did not overexpress NTN1 in Aβ-treated cell, previous studies have demonstrated the relevance of NTN1 in Aβ-related pathology72,73. Nevertheless, we anticipate the effect of NTN1 in Aβ-induced cells; because, NTN1 expression is reduced in AD brains, and its deficiency has been linked to apoptosis and neuronal cell death76,77. Given that cognitive impairment typically arises following the accumulation of Aβ and tau proteins, we anticipate that miR-937-3p-I could be developed as a therapeutic agent for early, preclinical AD by specifically targeting Aβ pathology78.

In conclusion, our study discovered a novel function of miR-937-3p-I in AD progression and elucidated the mechanisms involving NTN1 and apoptosis signaling. These insights highlighted the potential of miR-937-3p/NTN1/apoptosis signaling as a therapeutic target in AD treatment.