Knockdown of long non-coding RNA SOX21-AS1 attenuates amyloid-β-induced neuronal damage by sponging miR-107

Abstract Background: Alzheimer’s disease (AD), which has no effective drugs to delay or prevent its progression, is a multifactorial complex neurodegenerative disease. Long non-coding RNA SOX21 antisense RNA1 (SOX21-AS1) is associated with the development of AD, but the underlying molecular mechanism of SOX21-AS1 in AD is still largely unclear. Methods: To construct the AD model, SH-SY5Y and SK-N-SH cells were treated with amyloid-β1-42 (Aβ1-42). Quantitative real-time polymerase chain reaction (qRT-PCR) was executed to detect the expression of SOX21-AS1 and miRNA-107. Western blot analysis was utilized to assess the levels of phosphorylated Tau (p-Tau). 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) or flow cytometry assay was employed to determine the viability and apoptosis of SH-SY5Y and SK-N-SH cells. The relationship between SOX21-AS1 and miRNA-107 was verified with the dual-luciferase reporter assay. Results: SOX21-AS1 expression was augmented while miR-107 expression was decreased in Aβ1-42-treated SH-SY5Y and SK-N-SH cells. Moreover, Aβ1-42 elevated the levels of p-Tau and impeded viability and induced apoptosis of SH-SY5Y and SK-N-SH cells. Also, SOX21-AS1 silencing attenuated Aβ1-42 mediated the levels of p-Tau, viability, and apoptosis of SH-SY5Y and SK-N-SH cells. Importantly, SOX21-AS1 acted as a sponge for miR-107 in SH-SY5Y and SK-N-SH cells. Furthermore, the increase in p-Tau levels and apoptosis and the repression of viability of Aβ1-42-treated SH-SY5Y and SK-N-SH cells mediated by miR-107 inhibition were partly recovered by SOX21-AS1 depletion. Conclusion: SOX21-AS1 silencing could attenuate Aβ1-42-induced neuronal damage by sponging miR-107, which provided a possible strategy for the treatment of AD.


Introduction
Alzheimer's disease (AD) is a complex neurodegenerative disorder characterized by memory loss, cognitive and functional abilities decline that can lead to dementia [1]. The emphatic pathological features of AD are extracellular amyloid-β peptide (Aβ) deposition and intracellular neurofibrillary tangles formation [2]. Aβ is a neurotoxic agent that can couple with neurons and trigger cell death under certain aggregation conditions [3]. However, Aβ is a product of metabolism in the body, and its excessive deposition and clearance imbalance can lead to the occurrence of AD [4]. Neurofibrillary tangles are composed of highly p-Tau protein, and its accumulation is deemed to be closely connected to cognitive recession of AD [5]. At present, there are no effective treatments to delay or prevent the development of AD. Therefore, it is essential to explore the mechanism of the occurrence and development of AD to develop effective AD treatment strategies.
Long non-coding RNA (lncRNAs) are class of transcripts over 200 nucleotides in length that cannot be translated into proteins and exert a vital role in human physiology and disease processes [6]. There have been reported to be involved in transmission of genetic information, such as transcription factor recruitment, alternative splicing control, and mRNA stability and translational activity [7]. A substantial amount of evidence indicated that lncRNAs were involved in pluripotency maintenance, euro-ectodermal differentiation, epigenetic imprinted neuron-specific relaxation, brain tissue patterning and synaptogenesis, and mechanisms of epigenetic underlie memory formation [8]. Long non-coding RNA SOX21 antisense RNA1 (SOX21-AS1) is a recently discovered lncRNA, located on chromosome 13q32.1, and transcribed into a transcript of 2986 nucleotides [9]. Recent literature had revealed that SOX21-AS1 could reduce oxidative stress and inhibit neuronal apoptosis in AD mice [10]. However, the underlying molecular mechanisms of SOX21-AS1 in the pathogenesis of AD are rarely reported.
MicroRNAs (miRNAs) is a class of non-coding RNAs with approximately 21 nucleotides, which can be used as an important molecular tool for human disease diagnosis and prognosis evaluation [11]. They could regulate a variety of biological processes, containing neuronal differentiation, regeneration and survival [12]. Studies had shown that a large amount of miRNAs were involved in the progression of neurodegenerative disorder [13,14]. MicroRNA-107 (miR-107) was reported to be down-regulated in AD [15]. To date, it is unclear whether miR-107 is regulated by SOX21-AS1 and participates in the process of AD.
In the present study, we probed into the role of SOX21-AS1 in the pathogenesis of AD and its underlying molecular mechanism by constructing an AD cell model, and provided a possible strategy for the prevention and treatment of AD.

Materials and methods
Cell culture and treatment SH-SY5Y and SK-N-SH cells were obtained from the American Type Culture Collection (ATCC, Manassas, VA, U.S.A.). All cells were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum (Gibco, Thermo Fisher Scientific, Waltham, MA, U.S.A.). The above cells were maintained in an incubator in a humidified with 5% CO 2 at 37 • C. Aβ 1-42 was purchased from Sigma-Aldrich (St. Louis, MO, U.S.A.) for the construction of the AD cell model. First, in order to aggregate Aβ 1-42 , 0.5 mg Aβ 1-42 was dissolved in double distilled water at 37 • C for 7 d. Then, the thioflavin T assay was used to assess the aggregation of Aβ 1-42 . Next, the aggregation of Aβ 1-42 was diluted to a concentration of 200 μM and stored at −80 • C. SH-SY5Y and SK-N-SH cells were treated with different concentrations of Aβ 1-42 (0, 5, 10, 20 μM) for 24 h or Aβ 1-42 (10 μM) for different times (0, 12, 24, 48 h).

Flow cytometry assay
The apoptosis of SH-SY5Y and SK-N-SH cells with or without Aβ 1-42 treatment were detected by an annexin V-FITC assay kits (Invitrogen). In brief, treated cells were harvested and re-suspended with binding buffer (1×). Then, annexin V-FITC (5 μl) and propidium iodide (10 μl) were added to the binding buffer and incubated for 20 min in the dark. The apoptotic rate was determined by a FACScan ® flow cytometry (BD Biosciences, San Jose, CA, U.S.A.).

Dual-luciferase reporter assay
The binding sites between SOX21-AS1 and miR-107 were predicted with LncBase v.2 database. The sequences of SOX21-AS1 (with predicted miR-107 binding sites) and mutant SOX21-AS1 were cloned into the pGL3-control luciferase reporter vectors (Promega, Madison, WI, U.S.A.) for the construction of the luciferase reporter vectors SOX21-AS1-WT and SOX21-AS1-MUT. Then, the luciferase reporter vectors were co-transfected with miR-107 or miR-NC into SH-SY5Y and SK-N-SH cells using Lipofectamine 2000 transfection reagent (Invitrogen), respectively. The luciferase activity of the luciferase reporter vectors was assessed using dual-luciferase Reporter assay kit (Promega).

RNA immunoprecipitation (RIP) assay
EZ-Magna RIP RNA-Binding Protein Immunoprecipitation Kit (Millipore) was utilized to verify the relationship between SOX21-AS1 and miR-107 based on the manufacture's protocol. Briefly, cells were transfected with miR-107 or miR-NC for 48 h and then lysed in RIP buffer. Subsequently, the lysates of cells were maintained in a RIP buffer containing magnetic beads conjugated that are coupled with Ago2 or IgG antibody (Millipore). The enrichment of SOX21-AS1 was assessed by qRT-PCR.

Statistical analysis
The data in the study were derived from at least three independent experiments. Statistical analysis was conducted by SPSS 17.0 (SPSS, Chicago, IL, U.S.A.), and P < 0.05 was deemed to indicate a statistically significant difference. The differences between two or among more groups in the present study were determined using Student's t test or one-way analysis of variance followed by Tukey post-hoc test, respectively. Data from repeated experiments were presented as mean + − standard deviation.

Knockdown of SOX21-AS1 attenuated Aβ 1-42 mediated viability and apoptosis of SH-SY5Y and SK-N-SH cells
In consideration of the above findings, SH-SY5Y and SK-N-SH cells treated with Aβ 1-42 (10 μM) for 24 h were selected for further studies. First, Aβ 1-42 -treated SH-SY5Y and SK-N-SH cells were transfected with si-NC or si-SOX21-AS1 to silence the expression of SOX21-AS1. As shown in Figure 2A,B, the enhancement of SOX21-AS1 in SH-SY5Y and SK-N-SH cells induced by Aβ 1-42 was partly restored by the knockdown of SOX21-AS1, indicating that the si-SOX21-AS1 could be used for subsequent analysis. Afterward, the effects of SOX21-AS1 suppression on the viability and apoptosis of Aβ 1-42 -treated SH-SY5Y and SK-N-SH cells were determined by MTT or flow cytometry assays. Results of the MTT assay indicated that the inhibition of viability of SH-SY5Y and SK-N-SH cells induced by Aβ 1-42 was partially reversed by SOX21-AS1 silencing ( Figure 2C,D). Besides, flow cytometry assay revealed that SOX21-AS1 knockdown partially restored the facilitation of apoptosis of SH-SY5Y and SK-N-SH cells mediated by Aβ 1-42 ( Figure 2E,F). Also, Aβ 1-42 constrained the expression of Bcl-2 protein and elevated the expression of Bax in SH-SY5Y and SK-N-SH cells, while this influence was overturned by SOX21-AS1 suppression ( Figure 2G,H). Taken together, knockdown of SOX21-AS1 could recover the viability and apoptosis of SH-SY5Y and SK-N-SH cells induced by Aβ 1-42 .

SOX21-AS1 inhibition overturned Aβ 1-42 mediated the levels of p-Tau of SH-SY5Y and SK-N-SH cells
Subsequently, the influence of SOX21-AS1 silencing on the levels of p-Tau of Aβ 1-42 -treated SH-SY5Y and SK-N-SH cells was assessed by Western blot analysis. The results presented that Aβ 1-42 remarkably increased the level of p-Tau in SH-SY5Y and SK-N-SH cells, while this enhancement was overturned by the introduction of si-SOX21-AS1 ( Figure  3A-D). These findings indicated that SOX21-AS1 down-regulation could undermine Aβ 1-42 mediated the levels of p-Tau in SH-SY5Y and SK-N-SH cells.

SOX21-AS1 served as a sponge for miR-107 in SH-SY5Y and SK-N-SH cells
To further explore the underlying molecular mechanism of SOX21-AS1 in pathogenesis of AD, the LncBase v.2 database was utilized for the prediction of the potential target of SOX21-AS1. Bioinformatics analysis presented that miR-107 possessed a complementary base-paring region with SOX21-AS1 ( Figure 4A). To verify this prediction, the luciferase vectors SOX21-AS1-WT and SOX21-AS1-MUT were constructed. Results of dual-luciferase reporter assay displayed that the luciferase activity of the luciferase reporter vectors SOX21-AS1-WT in SH-SY5Y and SK-N-SH cells transfected with miR-107 was remarkably reduced than the control group, whereas the luciferase activity of the luciferase reporter vectors SOX21-AS1-MUT did not changed ( Figure 4B,C). Then, RIP assay was performed for further analysis of the relationship between SOX21-AS1 and miR-107. The results indicated that the enrichment of SOX21-AS1 in the Ago2-immunoprecipitation complex in SH-SY5Y and SK-N-SH cells transfected with miR-107 was markedly higher than that in the miR-NC group ( Figure 4D,E). In sum, these results manifested that SOX21-AS1 acted as a sponge for miR-107 in SH-SY5Y and SK-N-SH cells.

Knockdown of SOX21-AS1 reversed miR-107 inhibition-induced viability and apoptosis of Aβ 1-42 -treated SH-SY5Y and SK-N-SH cells
Next, we further investigated whether SOX21-AS1 affects the viability and apoptosis through miR-107 in Aβ 1-42 -treated SH-SY5Y and SK-N-SH cells. As exhibited in Figure 6A,B, miR-107 expression was dramatically decreased in Aβ 1-42 -treated SH-SY5Y and SK-N-SH cells transfected with anti-miR-107 compared with the anti-miR-NC group, while this reduction was reversed by SOX21-AS1 inhibition. MTT assay showed that the repression of viability of Aβ 1-42 -treated SH-SY5Y and SK-N-SH cells caused by miR-107 inhibition was abolished by the knockdown of SOX21-AS1 ( Figure 6C,D). In addition, flow cytometry assay revealed that inhibition of miR-107 markedly accelerated the apoptosis of Aβ 1-42 -treated SH-SY5Y and SK-N-SH cells, while this effect was abolished by SOX21-AS1 suppression ( Figure 6E,F). Furthermore, both the inhibition of Bcl-2 protein and the acceleration of Bax protein in Aβ 1-42 -treated SH-SY5Y and SK-N-SH cells mediated by miR-107 silencing were restored by SOX21-AS1 depletion ( Figure 6G,H). Therefore, these results indicated that SOX21-AS1 mediated the viability and apoptosis of Aβ 1-42 -treated SH-SY5Y and SK-N-SH cells via miR-107.

SOX21-AS1 repression weakened miR-107 inhibition-induced the levels of p-Tau of Aβ 1-42 -treated SH-SY5Y and SK-N-SH cells
Considering that the above findings, we further probed whether SOX21-AS1 affects the levels of p-Tau by miR-107 in Aβ 1-42 -treated SH-SY5Y and SK-N-SH cells. Western blot analysis manifested that silenced miR-107 expression could markedly elevate the level of p-Tau in Aβ 1-42 -treated SH-SY5Y and SK-N-SH cells, while this increase was recovered by SOX21-AS1 knockdown ( Figure 7A-D). Collectively, these data disclosed that SOX21-AS1 regulated the levels of p-Tau by miR-107 in Aβ 1-42 -treated SH-SY5Y and SK-N-SH cells.

Discussion
AD is a neurodegenerative disease associated with the elderly, whose pathology shows obvious deposition of Aβ and formation of neurofibrillary tangles. Aβ can self-aggregate to form oligomers which are toxic to the neuronal cells [16]. Studies had shown that Aβ induced neurotoxicity similar to the pathology of AD, and was widely used to construct AD model in vitro [17]. It was reported that lncRNAs were associated with neurological diseases, such as neurodegenerative disorder and epilepsy [18,19]. Studies had revealed that the hyper-phosphorylation of Tau leads to the formation of neurofibrillary tangles, which is highly related to the neurodegenerative changes in AD [20,21]. In the present study, we constructed the AD cell model in vitro to explore the role and potential molecular mechanism of SOX21-AS1 in the pathogenesis of AD.
Previous studies pointed out that lncRNAs exist curial roles in the pathogenesis of AD. For instance, knockdown of lncRNA BACE1-AS could enhance memory and learning behaviors in AD animal model [22]. Zhao et al. disclosed that lncRNA NEAT1 promoted the progression of AD by modulating the miR-124/BACE1 axis [23]. Also, reduced lncRNA EBF3-AS expression could overturn Aβ-induced apoptosis of SH-SY5Y cells [24]. In the present study, Aβ could enhance p-Tau protein level, repressed viability and promoted apoptosis of SH-SY5Y and SK-N-SH cells. Also, SOX21-AS1 was up-regulated in Aβ 1-42 -treated SH-SY5Y and SK-N-SH cells. Moreover, SOX21-AS1 inhibition could restore Aβ-mediated p-Tau levels, viability, and apoptosis of SH-SY5Y and SK-N-SH cells. Report of Zhang et al. reported that SOX21-AS1 silencing could suppress apoptosis of hippocampal neuron cells and enhance memory and learning ability in AD mice [10]. These findings manifested that SOX21-AS1 acted as a pathogenic factor in AD, and which was in line with the report of Zhang et al.
LncRNA had been shown to be involved in gene regulation as sponges for miRNAs or as endogenous competitive RNA [25]. Hu et al. suggested that miR-107 down-regulation could repress SYK expression and inactivate the NF-KB signaling pathway, which promoted the deterioration of spatial memory [26]. A previous report revealed that miR-107 family silencing might be played a vital role in AD pathogenesis via enhancing CDK5 activity [27]. Shu et al. claimed that miR-107 mimics overturned the enhancement of p-Tau induced by Aβ in mice [28]. In the present study, we found that SOX21-AS1 served as a sponge for miR-107 in SH-SY5Y and SK-N-SH cells. Besides, miR-107 was decreased in Aβ 1-42 -treated SH-SY5Y and SK-N-SH cells, and SOX21-AS1 repression reversed miR-107 inhibition-mediated p-Tau levels, viability, and apoptosis of Aβ 1-42 -treated SH-SY5Y and SK-N-SH cells. These results concluded that SOX21-AS1 mediated Aβ-induced neuronal damage via miR-107.