MicroRNA-144 Is Regulated by Activator Protein-1 (AP-1) and Decreases Expression of Alzheimer Disease-related A Disintegrin and Metalloprotease 10 (ADAM10)*

Background: MicroRNA (miR) dysregulation is found in Alzheimer disease (AD). A disintegrin and metalloprotease 10 (ADAM10) prevents generation of amyloid β (Aβ) and decrease AD pathology. Results: miR-144 suppresses ADAM10 expression and is up-regulated by activator protein-1. Conclusion: miR-144 is a negative regulator of ADAM10 and may be involved in AD pathogenesis. Significance: The first work to demonstrate the function of miRNA-144 and its regulation in the pathogenesis of AD. Amyloid β-peptide (Aβ) accumulating in the brain of Alzheimer disease (AD) patients is believed to be the main pathophysiologcal cause of the disease. Proteolytic processing of the amyloid precursor protein by α-secretase ADAM10 (a disintegrin and metalloprotease 10) protects the brain from the production of the Aβ. Meanwhile, dysregulation or aberrant expression of microRNAs (miRNAs) has been widely documented in AD patients. In this study, we demonstrated that overexpression of miR-144, which was previously reported to be increased in elderly primate brains and AD patients, significantly decreased activity of the luciferase reporter containing the ADAM10 3′-untranslated region (3′-UTR) and suppressed the ADAM10 protein level, whereas the miR-144 inhibitor led to an increase of the luciferase activity. The negative regulation caused by miR-144 was strictly dependent on the binding of the miRNA to its recognition element in the ADAM10 3′-UTR. Moreover, we also showed that activator protein-1 regulates the transcription of miR-144 and the up-regulation of miR-144 at least partially induces the suppression of the ADAM10 protein in the presence of Aβ. In addition, we found that miR-451, a miRNA processed from a single gene locus with miR-144, is also involved in the regulation of ADAM10 expression. Taken together, our data therefore demonstrate miR-144/451 is a negative regulator of the ADAM10 protein and suggest a mechanistic role for miR-144/451 in AD pathogenesis.

significantly decreased activity of the luciferase reporter containing the ADAM10 3'-untranslated region (3'-UTR) and suppressed ADAM10 protein level, whereas miR-144 inhibitor led to an increase of the luciferase activity and the protein. The negative regulation caused by miR-144 was strictly dependent on the binding of the miRNA to its recognition element in the ADAM10 3'-UTR. Moreover, we also showed that Activator protein-1 regulates the transcription of miR-144 and the up-regulation of miR-144 at least partially induces the suppression of the ADAM10 protein in the presence of Aβ. In addition, we found that miR-451, a miRNA processed from a single gene locus with miR-144, is also involved in the regulation of ADAM10 expression. Taken together, our data therefore demonstrate miR-144/451 is negative regulators of ADAM10 protein and suggest a mechanistic role for miR-144/451 in AD pathogenesis.
Alzheimer's disease (AD) is a neurodegenerative disease that accounts for 50 to 60% of dementia cases in the elderly (1). The amyloid β-peptide (Αβ) is believed to be the essential cause of AD pathology and is produced by sequential proteolytic cleavages of the amyloid precursor protein (APP) by βand γ-secretases (2). APP is also processed in a non-amyloidogenic pathway by α-secretase, thereby repressing Αβ formation (3). This alternative processing of APP by α-secretase generates the neuroprotective and neurotrophic soluble APPsα ectodomain (4). ADAM10, whose level is decreased in the platelets and neurons of AD patients (5,6), is the most recognized candidate for α-secretase in cell culture and mouse models (7)(8)(9). Additionally, APPsα is also decreased in the cerebrospinal fluid (CSF) of AD patients (10,11), indicating a reduced amount and/or activity of ADAM10. A previous study demonstrates that neuronal over-expression of ADAM10 in APP transgenic mice significantly decreases Aβ generation, amyloid plaque load, and AD pathology, while over-expression of a dominant negative ADAM10 variant enhances Aβ formation (3). These reports shed light on the important role of ADAM10 in Aβ production and the pathogenesis of AD. Thus, studying the regulation of ADAM10 expression may help to delineate the underlying pathophysiology of AD and offer new therapeutic targets for the disease.
In this study, we used a combination of bioinformatics and experimental techniques to demonstrate that miR-144 is a negative regulator of ADAM10, and Activator protein-1 (AP-1), that can be activated by Αβ and is implicated in AD pathogenesis, is involved in the regulation of miR-144 expression. In the presence of Αβ, miR-144 is up-regulated and contributes to the ADAM10 down-regulation caused by Αβ. Thus, this study presents a mechanism for the ADAM10 protein repression observed in AD patients and provides potential therapeutic targets for the prevention and/or treatment of the disease.

EXPERIMENTAL PROCEDURES
Bioinformatics -Prediction of miRNA targets was conducted using miRanda (http://www.microrna.org/). Prediction of the transcriptional start sites (TSSs) of primary microRNAs was performed using miRStart (http://mirstart.mbc.nctu.edu.tw/). The putative promoter sequences of primary microRNAs and the 3'UTR of ADAM10 were retrieved from the National Center for Biotechnology  Luciferase Reporter Gene Constructs-To construct reporter plasmid pmirGLO-ADAM10 3'UTR, the 3'-UTR region of ADAM10 was amplified from human cDNA. The PCR product was digested with SacI and XhoI. Then the fragment, which spans 1267 bps starting from the 35th bp upstream of the stop codon, was cloned into the pmirGLO Dual-Luciferase miRNA Target Expression Vector (Promega). To construct the luciferase reporter plasmids pGL3-promoter-2808, promoter-2554, promoter-1895, promoter-1019, promoter-453 and promoter-307, each promoter region of miR-144 was amplified from human genomic DNA and subsequently cloned into KpnI and XhoI site of pGL3-basic plasmid (Promega). The transcription factor over-expression constructs were amplified from human cDNA and then cloned into pcDNA3.1/myc-His A at the KpnI and XhoI sites or BamHI and XbaI sites. The ADAM10 3'-UTR luciferase mutant construct that lacks the putative miR-144 MRE, the ADAM10 3'-UTR luciferase mutant construct that lacks the putative miR-451 MRE and the miR-144 promoter reporter plasmids containing single or double site-specific deletion for each transcription factor AP-1 binding site were generated by using a PCR-mediated deletion mutagenesis protocol (31). The sequences of primers used in aforementioned construction or mutation are listed in Table 1 The primers used for quantitative RT-PCR are as follows: ADAM10 forward, 5'-TCGAACCATCACCCTGCAACCT-3' and reverse, 5'-GCCCACCAATGAGCCACAATCC-3'; and β-actin forward, 5'-GTCACCAACTGGGACGACATG-3' and reverse, 5'-GATAGCACAGCCTGGATAGCAAC-3'. β-actin served as an internal control. miRNAs were extracted using an miRcute miRNA isolation kit (TIANGEN). Extracted miRNAs were polyadenylated by poly(A) polymerase and reverse transcribed into cDNA using oligo-dT primers with an miRcute miRNA first-strand cDNA synthesis kit (TIANGEN) according to the manufacturer's protocol. Quantitative RT-PCR analysis for mature miR-144 was miR-144 Suppresses ADAM10 Expression performed with an miRcute miRNA qPCR detection kit (TIANGEN) using the primers miR-144 forward, 5'-GGGGGGGGGGGGGTACAGTATAGATG ATGTACTAA-3'; miR-451 forward, 5'-GGGGGGGGCCAAACCGTTACCATTAC TGAGTTAAAA-3'; and U6 forward, 5'-GCAAGGATGACACGCAAATTCGT-3'. The reverse primer for the quantitative RT-PCR was provided in the miRcute miRNA qPCR detection kit. U6 served as an internal control.
Electrophoretic Mobility Shift Assay (EMSA)-Nuclear proteins from SH-SY5Y cells were extracted by NE-PER Nuclear and Cytoplasmic Extraction Reagents (Pierce). The following biotin-labeled probes and unlabeled competitor probes with the same sequences as well as their anti-sense oligos were synthesized at Invitrogen and annealed: 5'-CTTCTAGGGAAAGGGGCCAGTGACCCT TGTCATGGACTCTAGCAGGGC-3', corresponding to AP-1 binding site 3 of miR-144 promoter; 5'-CCATAACCCACCTGGGCTGTGCCTGAC CACAGAATCAAGGAGACGCTG-3', corresponding to AP-1 binding site 5 of miR-144 promoter. EMSA reactions (20 μL) were assembled using LightShift Chemiluminescent EMSA Kit (Pierce), and the reaction contained 50 fmol of biotin-labeled probe, 2 μL 10× binding buffer, 1 μL l μg/μL Poly (dI• dC), 0.5 μL 50% glycerol, and/or 2μL NE-PER nuclear extracts, 10 pmol unlabeled probe. Mixtures of protein and DNA were incubated for 20 min at room temperature. Following addition of 5 μL loading buffer, bound and free DNA were resolved by fractionation on 6% native polyacrylamide TBE gel (Bio-Rad) in 0.5× TBE. Gels were pre-run for 40 min at 100 V, and run with samples under the same conditions for 40 min. The DNA/protein complex was transferred onto Biodyne B Nylon Membrane (PALL) at 380 mA for 1 h. The transferred DNA was cross-linked to the membrane at 120 mJ/cm 2 for 60 s using UV-light crosslinking instrument equipped with 254 nm bulbs (DETIANYOU Technology). The biotin-labeled DNA was detected using the Chemiluminescent Nucleic Acid Detection Module (Pierce).
Statistical Analysis-Results are presented as the mean of three independent experiments ± miR-144 Suppresses ADAM10 Expression the standard error of the mean (SEM). An independent two-tailed Student's t-test was performed. Differences were considered statistically significant at p < 0.05.

RESULTS
MiR-144 Regulates ADAM10 Negatively-The computational program Miranda was used to detect potential miRNAs that could bind to the 3'-UTR of human ADAM10 mRNA (NM_001110). Combined with previous reports, we picked up three miRNAs as potential candidates: miR-9, miR-101 and miR-144, which are implicated in amyloidogenic gene and/or aberrantly express in AD patients (25,29,32). The sequences of the candidate miRNAs and their predicted MREs in the ADAM10 3'-UTR are listed in Table 2. To further investigate the potential function of the candidate miRNAs on ADAM10 expression, luciferase assays were performed in human neuroblastoma SH-SY5Y cells. The result demonstrated that, except for the positive control miR-122 (33,34), only transfection with the miR-144 mimics resulted in a significant luciferase activity down-regulation of the ADAM10 3'-UTR reporter (Fig. 1A). A down-regulation in the luciferase activity after miR-144 transfection was also observed in HeLa cells (data not shown).
Previous genome-wide analysis of miRNA expression demonstrated that compared to global miRNA down-regulation, miR-144 is significantly increased both in elderly primate brains and in AD patients (32). Moreover, bioinformatics analysis revealed that the putative MRE for miR-144 in the ADAM10 3'-UTR is strictly conserved in vertebrates (Table 3). These observations further prompted us to analyze the function of miR-144 on ADAM10 expression. In the miR-144 loss-of-function experiments, we found that miR-144 inhibitor up-regulated the luciferase activity of the ADAM10 3'-UTR reporter (Fig. 1B). Moreover, ADAM10 protein levels are remarkably decreased upon transient over-expression of miR-144 in SH-SY5Y cells ( Fig. 1C and D). However, the ADAM10 protein levels were not altered after over-expression of miR-9, which also possesses a putative MRE in the ADAM10 3'-UTR, indicating the specificity of miR-144 in the regulation of ADAM10 levels. Identical data were obtained in HeLa cells (data not shown). In addition, a decrease in the mRNA level of ADAM10 upon miR-144 over-expression was not observed (Fig. 1E). These results thus identify miR-144 as a negative regulator of ADAM10 and suggest that the repression of the ADAM10 protein induced by miR-144 does not occur via ADAM10 mRNA degradation but may be achieved at the translational level.
Analysis of MiR-144 MRE within ADAM10 3'-UTR-Regulation of target genes by miRNAs is dependent on imperfect base pairing of the miRNAs to the 3'-UTR of their target mRNAs through a 2-to 8-nt seed region located near the 5'-end of the miRNA (35). As shown in Fig. 2A, miR-144 was predicted to possess a putative MRE within the ADAM10 3'-UTR. To investigate whether the miR-144-induced ADAM10 repression occurred via specific binding to the MRE in the ADAM10 mRNA 3'-UTR, a mutant luciferase reporter construct that lacks the potential MRE of miR-144 was generated. The luciferase assay analysis demonstrated that the mutant reporter construct was insensitive to miR-144-mediated inhibition compared to the wild type (Fig. 2B). This result indicates that the miR-144-induced ADAM10 repression depends on the direct binding of miR-144 to its MRE within the ADAM10 mRNA 3'-UTR.
Identification of MiR-144 Promoter-The demonstration that miR-144 is a negative regulator of ADAM10 expression prompted us to examine the transcriptional regulation of miR-144 and its implication(s) in AD. Using the publicly accessible algorithm MiRStart (36), which integrates three datasets including cap by guest on October 5, 2017 http://www.jbc.org/ analysis of gene expression (CAGE) Tags (37,38), transcriptional start site (TSS) Seq libraries and H3K4me3 chromatin signatures, the site at -371 bp upstream of the miR-144 precursor was predicted as the TSS of the primary miR-144. To experimentally verify this putative TSS, we generated a luciferase reporter construct containing the 2.8-kb region upstream of the precursor miR-144, named pGL3-promoter-2808. According to sequence conservation data, five additional luciferase reporter constructs containing decreasing lengths of the upstream sequence (-2554, -1895, -1019, -453, and -307) were derived. The six luciferase reporter constructs and the pGL3-basic plasmid were then each transfected into SH-SY5Y cells to determine the basal promoter activity. As shown in Fig. 3, the pGL3-promoter-453 construct (i.e., the 453-bp putative promoter region) displayed the highest luciferase activity, which was approximately 10-fold greater than pGL3-basic. The fact that the other four reporter constructs with longer upstream sequences exhibited luciferase activities at comparable levels to that of pGL3-promoter-453 and that pGL3-promoter-307 displayed a dramatic decrease in luciferase activity further support that the TSS is located in the -307 to -453-bp region. The similar result was also observed in HeLa cells (data not shown). These results match our bioinformatics prediction and indicate that the 453-bp region upstream of the precursor miR-144 possesses intact promoter activity triggering miR-144 transcription.
Transcriptional Regulation of MiR-144-To investigate the key transcription factors involved in miR-144 transcription, the ~1-kb region upstream of the miR-144 TSS was analyzed by transcription factor binding prediction software. SP1, AP-1, CREB, CP2, and TCF4 were predicted to possess putative binding sites in this region. Thus, these five transcription factor were cloned into expression vectors and each co-transfected with the pGL3-promoter-1019 reporter into SH-SY5Y cells. As shown in Fig. 4A, relative to the other four transcription factors, over-expression of c-Jun, a common component of the AP-1 protein complex and can bind to AP-1 recognition sites (39)(40)(41), resulted in a strong increase in the luciferase activity. Moreover, the luciferase activity was also increased in a dose-dependent manner after treatment of TPA, which can induce transcription from AP-1-driven promoter (42,43), (Fig. 4B). Meanwhile, the endogenous expression of miR-144 was also significantly increased by c-Jun over expression (Fig. 4C) and TPA treatment (Fig. 4D), but decreased after endogenous c-Jun was knocked down (Fig. 4E).
As shown in Fig. 4F, six potential AP-1 binding sites are predicted to locate in the ~1-kb region upstream of the miR-144 precursor. We used deletion mutagenesis to abolish each site and then performed luciferase assays. The results showed that, compared to the wild type, the mut-3, mut-5 and the double mutation (mutation of binding sites 3 and 5) constructs inhibited the TPA-induced transcriptional activity by 26.1%, 26.4% and 41.6% respectively (Fig. 4G). However, mutation of any other AP-1 binding site did not remarkably affect the transcriptional activity upon TPA treatment. The similar result was observed when c-Jun was over-expressed (Fig. 4H). In addition, the binding properties of the AP-1 at the sites 3 and 5 were further validated by performing EMSA and supershift assay. DNA probes corresponding to AP-1 binding sites 3 and 5, which contained the putative AP-1 binding sequences and their flanking region at both sides in the human miR-144 promoter, were tested. Incubation of either biotin-labeled probe with SH-SY5Y cell nuclear extracts caused a major retarded band (Fig. 4I, lane 3 and 7). Excessive amount of unlabeled probes prevented each labeled probe from binding with AP-1, as indicated by the dim retarded band (Fig. 4I, lane 4 and 8). Additionally, incubation of c-Jun antibody by guest on October 5, 2017 http://www.jbc.org/ caused supershift bands (Fig. 4I, lane 2 and 6), indicating the presence of c-Jun in the binding complex. These results fully demonstrate that AP-1 can regulate the transcription of miR-144 by binding the potential sites in its promoter.
MiR-144 Contributes to the Down-regulation in ADAM10 Protein in the Presence of TPA -In vivo evidence demonstrates that the protein levels of ADAM10 and/or APPsα are decreased in AD patients (5,10,11), whereas expression or activity of AP-1/c-Jun is increased (44)(45)(46). Additionally, we have demonstrated that miR-144 acts as a negative regulator of ADAM10 and is regulated by AP-1/c-Jun. Prompted by these observation, we next investigated whether the ADAM10 changes in the presence of TPA, and whether miR-144 might be causally involved in regulating such changes. To test this, we initially determined both ADAM10 mRNA and protein levels in SH-SY5Y cells after TPA treatment. We found that, although the level of ADAM10 mRNA was unchanged in the presence of TPA (Fig. 5A), the level of ADAM10 protein was down-regulated sustainedly in the first 24 hrs and then started to increase ( Fig. 5B and C). The inconsistent changes between the mRNA and protein strongly imply that the down-regulation of ADAM10 protein in the presence of TPA is primarily exerted at the translational level. Meanwhile, using quantitative RT-PCR analysis, we found that the endogenous mature miR-144 was induced after TPA treatment, peaked at 12 hrs, and then started to decrease (Fig. 5D). Notably, the increase in the level of miR-144 exhibited an evident inverse correlation with the change of the ADAM10 protein in at least the first 12 hrs after the initiation of the TPA treatment (Fig. 5E). The incomplete inverse correlation displayed after the 12 h point may be due to involvement of other signal pathways activated by TPA. Furthermore, miR-144 loss-of-function experiments revealed that the repression of ADAM10 protein induced by TPA was restored by miR-144 inhibition ( Fig. 5F and G). These data thus further supports the notion that miR-144 is a negative regulator of ADAM10 protein and strongly imply that the AD-associated increase in expression or activity of AP-1/c-Jun as well as elevated expression in miR-144 may contribute to the reduction of ADAM10 protein in AD pathogenesis.
MiR-144 Contributes to ADAM10 Down-regulation upon Aβ Treatment-As one of the crucial pathological causes of AD, Aβ phosphorylates and activates c-Jun in vitro (47)(48)(49). Thus, we further investigated the levels of miR-144 in SH-SY5Y cells after incubation with Aβ 42 peptide, which is amyloidogenic and is the major component of neuritic plaques (50,51).
Quantitative RT-PCR analysis demonstrated that mature miR-144 levels were strongly increased after incubation with Aβ 42 peptide (Fig. 6A). Meanwhile, we analyzed ADAM10 protein and mRNA levels after Aβ 42 treatment. As shown in Fig. 6B and C, the ADAM10 protein level significantly decreased upon treatment with Aβ 42 . Additionally, APP carboxy-terminal fragment (α-CTF), a membrane-bound APP derivant truncated by ADAM10, was also suppressed, and APP β-CTF was correspondingly increased in the presence of Aβ 42 , further indicating a decrease in ADAM10 expression or activity. However, ADAM10 mRNA levels were not changed upon Aβ 42 treatment (Fig. 6D), implying that the Aβ 42 -induced ADAM10 down-regulation is also mainly achieved at the translational level. To further investigate whether the increase of miR-144 is involved in the down-regulation of ADAM10 protein caused by Aβ 42 , we performed a miR-144 loss-of-function experiment. As shown in Fig. 6E  protein levels. However, the fact that the miR-144 inhibitor did not completely abolish the decrease in ADAM10 induced by Aβ 42 implies that the regulation of ADAM10 expression cannot solely be explained by the action of miR-144 and that additional regulatory pathways must be involved.

MiR-451, A MiRNA Processed from A Single Gene Locus with MiR-144, Is Also Involved in The Regulation of ADAM10
Expression -MiR-451 is co-transcribed with miR-144 from a single locus (52) and is predicted to bind to the ADAM10 3'-UTR ( Table 2). In some cases, miR-451 and miR-144 share a same mechanism of transcriptional regulation and display synergistic depressive effects on their target proteins (52,53). Thus, we investigated whether miR-451 is involved in the regulation of ADAM10 protein and whether it is also controlled by c-Jun. We found that, although the putative MRE for miR-451 in the ADAM10 3'-UTR is not conserved in vertebrates (data not shown), over-expression of miR-451 in SH-SY5Y cells also attenuated ADAM10 protein (Fig. 7A and B), and didn't change ADAM10 mRNA level (Fig. 7C). Meanwhile, miR-451 over-expression significantly decreased the activity of the luciferase reporter containing the ADAM10 3'-UTR, but not that of a mutant luciferase reporter construct that lacks the potential miR-451 MRE (Fig. 7D), indicating that the ADAM10 repression caused by miR-451 also depends on the binding of MRE in ADAM10 3'-UTR. Similar to miR-144, the level of endogenous mature miR-451 was increased by c-Jun over-expression (Fig. 7E) and decreased when c-Jun was knocked down (Fig. 7F). Moreover, in the presence of TPA, the level of miR-451 also changed in a time-dependent manner (Fig. 7G), displayed a nearly identical pattern of miR-144 and also inversely correlated with ADAM10 protein levels in the first 12 hrs after the initiation of the TPA treatment (Fig.   7H). In addition, the level of mature miR-451 was also significantly increased after incubation with the Aβ 42 peptide (Fig. 7I). These results suggest that, as another negative regulator of ADAM10 protein, miR-451 may share a same function and mechanism of transcriptional regulation with miR-144 in AD pathogenesis.

DISCUSSION
Brain aging is a major risk factor for the development of AD and other prevalent neurodegenerative disorders (54). This implies that genetic changes occur in the process of brain aging that play coordinating roles with environmental stress to lead to the onset of AD and other neurodegenerative pathologies. However, these genetic changes and their implications in such diseases remain largely unknown. Genome-wide analysis of miRNA expression has revealed that compared to global miRNA down-regulation, miR-144 is the sole miRNA that is consistently elevated in the brains of elderly humans, chimpanzees, and rhesus macaques, and it is also increased in AD patients (32). Because miR-144 could suppress the expression of ADAM10, the elevated miR-144 in aging brains at least partially induces the depletion of ADAM10 protein levels and increases the susceptibility to environmental stress, which may in turn determine the onset of AD. Moreover, growing evidence supports the concept that AD is fundamentally a metabolic disease with derangements in brain glucose utilization and responsiveness to insulin and insulin-like growth factor (IGF) stimulation. The impaired insulin and IGF signaling contributes to AD-associated neuronal loss, synaptic disconnection, tau hyper-phosphorylation, and Aβ accumulation (55). Similarly, miR-144 is also increased and impairs insulin signaling through down-regulation of insulin receptor substrate 1 (IRS1) (56)  Recent reports document that ADAM10 expression is regulated both transcriptionally and post-transcriptionally. Specifically, retinoic acid enhances ADAM10 transcription by promoting the binding of a non-permissive dimer of retinoic acid receptor-α (RARα) and retinoid X receptor-β (RXRβ) to retinoic acid-responsive elements in the promoter of the ADAM10 gene (57,58). ADAM10 transcription is also regulated by PAX2 in renal cancer cells (59) and melanocytes (60). In addition, expression of ADAM10 is suppressed by its 5'-UTR (61), and ADAM10 activity can be blocked by its own predomain (62). Although many reports on ADAM10 regulation have been published in past several years, none present a reasonable mechanism for ADAM10 down-regulation in AD patients. Here, we demonstrated that miR-144 suppresses ADAM10 expression via a classical interaction with the ADAM10 3'-UTR. Thus, the elevated levels of miR-144 in AD patients may be responsible for the down-regulation of ADAM10 protein levels. Additionally, it was observed that the level of ADAM10 mRNA is not affected by miR-144 or TPA (Fig. 1E, 5A), suggesting that the negative regulation of ADAM10 by miR-144 is mainly achieved at the translational level. Same as our basic research finding, the clinical reports showed that ADAM10 proteins levels are decreased in the neurons of AD patients (6), but ADAM10 mRNA levels are increased in hippocampal samples from severe AD cases (63). These two seemingly incompatible observations also imply that translational regulation of ADAM10 exists in the pathogenesis of AD.
The functions of miRNAs in gene regulation have been extensively studied in the last decade. At the same time, research interests have gradually focused on the mechanism(s) by which miRNAs are controlled. In our study, we both identified miR-144 as a negative regulator of ADAM10 and demonstrated that AP-1/c-Jun regulates miR-144 transcription. Consistent with our results, recent studies also demonstrate that c-Jun regulates expression of miRNAs at transcriptional level (64)(65)(66). Moreover, there is evidence showing an increase in c-Jun expression in the neurons of AD patients (45,47). In postmortem pathological brain samples, phosphorylated c-Jun staining is found only in the affected regions of the cerebral cortex and hippocampus (67). Furthermore, c-Jun influences the initiation and execution of Aβ-induced neuronal apoptosis (68)(69)(70). These findings suggest that the induction of c-Jun is crucial to the pathogenesis of AD. However, previous studies defining the implication of the AP-1/c-Jun pathway in AD mainly concentrate on its involvement in neuronal apoptosis. Here, we demonstrated that AP-1 may directly participate in amyloidogenesis by regulating miR-144.
Notably, AP-1 is not the only transcription factor activated after Aβ treatment in vitro or Aβ deposition in vivo. For instance, there is evidence that Aβ treatment of human brain endothelial cells results in increases in the activities of the AP-1, CREB, GATA, NFATc, and GRE transcription factors (47). Among these, the GATA family was previously reported to regulate transcription of miR-144 and miR-451 (52,53). Thus, a vicious circle in AD pathogenesis composed by Aβ, AP-1/c-Jun, miR-144 and ADAM10 is emerging. In detail, miR-144 elevates in ordinary aging population or in early stage AD patients, and decreases ADAM10 protein levels, which hampers the non-amyloidogenic pathway, and forces metabolism of Aβ to shifts to favor Aβ production. When Aβ is over-produced, brain deterioration begin, then AP-1, GATA, and other transcription factors are recruited or activated, in turn, further promotes miR-144/451 expression and consequently reinforces the process (Fig. 8), which could be accelerate or interrupted by environmental factors. Although, till now, the miR-144 Suppresses ADAM10 Expression first participant who starts the vicious circle is not known yet, we can see that miR-144/451 play a crucial role in it.
In conclusion, this study provides evidence that miR-144 decreases the expression of ADAM10 protein and demonstrates that miR-144 is transcriptionally regulated by AP-1/c-Jun, suggesting miR-144 contributes to AD pathogenesis and pharmacological targeting miR-144 may represent a promising strategy in the management of AD.          (5 µM) was measured by quantitative RT-PCR and plotted as fold change with respect to PBS-treated control. B, ADAM10 protein, α-CTF, and β-CTF in SH-SY5Y cells 24 hrs after treatment with Aβ 42 peptide at the indicated concentrations were analyzed by western blot. C, Intensities of the ADAM10 bands from three independent experiments were quantified and normalized to that of corresponding GAPDH bands. The values were plotted as the fold change with respect to the PBS-treated control. D, Quantitative RT-PCR analysis of ADAM10 mRNA in SH-SY5Y cells 24 hrs after treatment with Aβ 42 peptide (5 µM). E, SH-SY5Y cells with or without miR-144 loss of function by transfection with the miR-144 inhibitor or mirRNAs inhibitor negative control (miR inhibitor NC) were treated with Aβ 42 (5 µM) peptide for 24 hrs, then the levels of ADAM10 protein in the cells were analyzed by western blot. F, Intensities of the ADAM10 bands from three independent experiments were quantified and normalized to that of corresponding GAPDH bands. The relative expression of ADAM10 was plotted as the fold change with respect to the negative control. Results in A, C, D and F are presented as means ± SEM from experiments performed in triplicate. Asterisks indicate significant differences (*p < 0.05).