MicroRNA-101 Regulates Amyloid Precursor Protein Expression in Hippocampal Neurons*

The amyloid precursor protein (APP) and its proteolytic product amyloid beta (Aβ) are associated with both familial and sporadic forms of Alzheimer disease (AD). Aberrant expression and function of microRNAs has been observed in AD. Here, we show that in rat hippocampal neurons cultured in vitro, the down-regulation of Argonaute-2, a key component of the RNA-induced silencing complex, produced an increase in APP levels. Using site-directed mutagenesis, a microRNA responsive element (RE) for miR-101 was identified in the 3′-untranslated region (UTR) of APP. The inhibition of endogenous miR-101 increased APP levels, whereas lentiviral-mediated miR-101 overexpression significantly reduced APP and Aβ load in hippocampal neurons. In addition, miR-101 contributed to the regulation of APP in response to the proinflammatory cytokine interleukin-1β (IL-lβ). Thus, miR-101 is a negative regulator of APP expression and affects the accumulation of Aβ, suggesting a possible role for miR-101 in neuropathological conditions.

Alzheimer disease (AD) 2 is the most common form of dementia in aged individuals and is characterized by A␤ plaques, which contain A␤ aggregates and neurofibrillary tangles which consist primarily of aggregated forms of the microtubule-stabilizing protein tau (reviewed in Ref. 1). A␤ peptides are derived from processing of the type I transmembrane protein APP through sequential cleavages by ␤ and ␥ secretase (2)(3). The A␤ load during pathology leads to neurological dysfunction. APP is linked to AD; familial AD can be caused by increased expression of APP due to either genomic duplication (4 -5) or regulatory sequence alterations (6). Among the physiological and pathological activa-tors of APP expression (7)(8) is the proinflammatory cytokine IL-1␤ (9). IL-1␤ is produced in the central nervous system (CNS) in response to damage and influences neuronal function by interacting with the type I IL-1 receptor expressed on neurons (10 -11). IL-1␤ is overexpressed in AD (12) and has been implicated in initiation and progression of AD pathology (13). In addition, IL-1␤ promotes APP transcription (14) and translation (9) in various cell types. Transcriptional and post-transcriptional regulation of APP expression has been widely studied and correlated to AD pathogenesis (15)(16). Both cell type-specific promoter elements (17) and regulatory elements in the 5Ј-and 3Ј-UTRs of APP mRNA have been identified (9,18).
MicroRNAs are an intriguing class of small noncoding RNA molecules which, in mammals, regulate gene expression primarily by imperfect base pairing with the 3Ј-UTR of specific target mRNAs (19). MicroRNAs associate with Argonaute proteins (Ago1-4, in mammals) (20 -21), which constitute the core of the RNA-induced silencing complex (RISC), and mediate post-transcriptional repression of target messenger RNAs (19). Ago2 is expressed at high levels in human (22) and mouse (23) brain and Argonaute expression profiling or depleting Ago1-4 has been used to identify potential microRNA targets (24 -25). Several studies have indicated that microRNAs define the spatial and temporal expression profiles of genes involved in neuronal development and differentiation (26). In addition, microRNAs are recruited during the execution of neuronal signal transduction pathways (27). Emerging evidence suggests that changes in expression of microRNAs are associated with neurodegenerative diseases (28). Profiling microRNAs from selected human brain areas has revealed significant changes in AD patients (29 -31). A few microRNAs, involved in the regulation of genes causally linked to Alzheimer's disease, are dysregulated in human AD patients (30,(32)(33) and AD mouse models (34).
The hippocampus is one of the main brain regions affected during the early stages of AD, and changes in the hippocampus coincide with the memory deficits observed in AD patients. Therefore, elucidation of the molecular mechanisms regulating APP expression in primary hippocampal neurons will be useful in understanding this disease.
In the present work, we focused on identifying a microRNA regulating APP expression in primary cultures of rat hippocampal neurons. In addition, we have analyzed the response of this microRNA to treatment of the cultures with IL-1␤.

EXPERIMENTAL PROCEDURES
Cell Cultures-Hippocampal neurons were prepared from embryonic day 17-18 (E17-18) embryos from timed pregnancy Wistar rats (Charles River). The hippocampus was dissected out in Hepes-buffered Hanks' balanced salt solution and dissociated via trypsin/EDTA treatment. Cells were plated at 1 ϫ 10 6 cells on 3.5-cm dishes precoated with poly(D/L-lysine) and cultured in neurobasal medium supplemented with B-27 and glutamax. Half of the medium was changed every 3-4 days. Rat PC12 cells were cultured in RPMI 1640 medium with 5% fetal calf serum and 10% heat-inactivated horse serum in a 5% CO 2humidified incubator at 37°C.
RNA Extraction and Analysis-Total RNA was extracted from cells using TRIzol (Invitrogen) according to the manufacturer's instructions. For Northern analysis of miRNAs, total RNA was separated on Tris-borate/EDTA 10% polyacrylamide 7 M urea denaturing gels and transferred to a Hybond-XL membrane (GE Healthcare). Hybridization was performed with 32 Plabeled DNA probes corresponding to U6 snRNA or miR-101 obtained by primer extension. In particular, tailed DNA probes were synthesized with the following oligonucleotides: universal template, TTTTTTTTTGGTAGG; miR-101a probe, TTC-AGTTATCACAGTACTGTACCATCC; U6 probe, AGTA-TATGTGCTGCCGAAGCCCATCC. Band intensities were calculated using the Quantity One (Bio-Rad) software and normalized to those obtained from U6 RNA. RNA quantitation was performed by real-time quantitative PCR. RNA was treated with a TURBO DNA-free TM kit (Ambion), reverse-transcribed with SuperScript III Reverse Transcriptase (Invitrogen), and amplified with FastStart Universal Probe Master (Roche) in a 7900HT Fast Real-Time PCR System (Applied Biosystems). Probes detecting TATAbinding protein (TBP), APP, and Ago2 were chosen from the Roche Universal Probe Library and used as recommended by the manufacturer with the corresponding primers designed by the QuikChange Primer Design Program. Probe 129 was used with rno-TBP-1029-1048FW 5Ј-CCCACCAGCAGT-TCAGTAGC-3Ј and rno-TBP-1081-1103 RW 5Ј-CAATTCTGG-GTTTGATCATTCTG-3Ј; probe 69 with rno-APP-659-678FW 5Ј-GGAGCGGACACAGACTATG-C-3Ј and rno-APP730 -750RW 5Ј-GCTTCTTCTTCCTCACATCG-3Ј; probe 77 with rno-AGO21548-1559FW 5Ј-AACACATACGCT-GGTCTCCA-3Ј and rno-1603 1621AGO2RW 5Ј-CTCCCACACG-CTTGACTTC-3Ј. Relative changes in gene expression were quantified using the comparative threshold method (Ct) after determining the Ct values for reference (TBP) and target genes in each sample set according to the 2 Ϫ⌬⌬Ct method. TBP was used as an endogenous control. All reactions were performed in triplicate.
Quantitative RT-PCR for miR-101-The Taqman microRNA reverse transcription kit (Applied Biosystems) and the FastStart Universal Probe Master (Roche) were used. The U6 snRNA was used for normalization of samples. The quantitative PCR procedure was carried out according to the instructions provided with the TaqMan microRNA assay kit (Applied Biosystems).
Immunocytochemistry-Hippocampal neurons were fixed for 15 min in phosphate-buffered saline containing 4% formaldehyde and 4% sucrose, permeabilized with 0.1% Triton X-100 for 8 min, and processed for labeling with mouse monoclonal anti-APP (4G8). Nuclei were stained with Hoechst 33258 (0.5 g/ml, Sigma Aldrich). A secondary antibody coupled to Alexa 594 was obtained from Molecular Probes (Invitrogen). Digital images were obtained with an Olympus BX51 microscope (100ϫ and 60ϫ oil objectives) equipped with a Spot Diagnostic Instruments camera and collected with Spot image analysis software.
Luciferase Reporter Gene Constructs and Luciferase Assay-The SG-APP 3Ј-UTR was purchased from Switch-Gear Genomics (SG-Luc-APP; Amplicon start: chr21:26174703; Amplicon end: chr21:26175879). The APP luciferase mutant constructs were generated using the QuikChange II Site-directed Mutagenesis kit (Stratagene) and synthetic oligonucleo-   . Inverse correlation between miR-101 and APP expression in hippocampus tissue. A, APP protein levels were analyzed by Western blotting of hippocampus from rats, P8 to 6 months of age, as indicated. The GAPDH signal was used as an internal control. Expression levels are shown relative to the P8 hippocampus. B, Northern blot analysis of mature miR-101 from total RNA. U6 RNA was used as control. Expression levels are shown relative to 6-month-old animals. C, the inverse correlation between APP protein and mature miR-101 levels is shown. Means from two independent experiments are shown. 9632, Sigma-Aldrich) and a control vector expressing an siRNA, which does not target rat mRNAs (SHC002, Sigma-Aldrich) were used. The following oligonucleotides, which contain the mature miR-101 sequence in an artificial stem loop were designed, annealed, and cloned into the pLB lentiviral vector (Addgene plasmid 11619) (36) using the HpaI and XhoI restriction sites: hsa-miR-101UP, 5Ј-TGTTCAGTT-ATCACAGTACTGTACTCG AGTACAGTACTGTGATAA-CTGAACTTTTTC; hsa-miR-101DOWN, TCGAGAAAAA-GTTCAGTTATCACAGTACTGTACTCGAGTA CAGTAC-TGTGATAACTGAACA. The following oligonucleotides were used to produce a pLB lentiviral vector expressing a control miRNA: shc002UP, TGCAACAAGA TGAAGAGCACC-AACTCGAGTTGGTGCTCTTCATCTTGTTGCTTTTTC; shc002DOWN, TCGAGAAAAAGCAACAAGATAAGAGCA CCAACTCGAGTTGGTGCTCTTCATCTTG TTGCA. The VSV-G pseudotyped lentiviral particles were generated by Fugene HD (Roche) transfection of HEK293T cells with a mixture of the pLB lentiviral vector and 3 plasmids (kindly provided by L. Naldini) (37) that are essential for the production of thirdgeneration lentiviruses, as previously described (38).
Statistical Analysis-Quantitative data are presented as the mean Ϯ S.D. Student's t test was used to determine significant differences between two groups.

Silencing of Ago2 in Hippocampal Neurons Increases APP
Protein-To explore the role of miRNAs in regulating APP gene expression, we first analyzed the level of APP protein in rat hippocampal cells in which the expression of Ago2 was downregulated. Hippocampal neurons were transduced with either a lentiviral vector containing Ago2 shRNA under the control of the U6 promoter or a control lentivirus expressing a scrambled siRNA. Ago2 siRNA expression resulted in a significant reduction of Ago2 mRNA in comparison to the control siRNA (Fig.  1A). Western blot analysis showed that APP levels were significantly higher in neurons in which Ago2 levels were reduced (Fig. 1, B and C). However, no significant alteration of APP mRNA was observed (Fig. 1D), suggesting that in hippocampal neurons APP translation may be regulated by an Ago2/ microRNA pathway.
miR-101 Is a Putative APP Regulator in the Rat Hippocampus-A combined data analysis using the computational programs TargetScan and Pictar for detecting microRNA responsive elements within the 3Ј-UTR of APP, which is conserved among the human, rat and murine genes, was conducted. Among the selected microRNAs ( Fig. 2A), miR-101 had two putative target sites within the APP gene ( Fig. 2A). Previous data indicate that mouse miR-101 was initially cloned as a brain-enriched microRNA (39) and in situ hybridization data from the Allen Brain Atlas have shown that the non-coding RNA AK021368, which encompasses microRNA 101a, is expressed in the mouse brain, including the hippocampus (40). These observations further prompted us to analyze APP and miR-101 expression during in vitro development of rat primary hippocampal neurons. In this cellular model, miR-101 levels increased during development from 2 up to 17 days in vitro (2B). In contrast, APP protein levels decreased (Fig. 2C), showing an inverse correla-

. Inhibition of miR-101 increases endogenous full-length APP.
A, hippocampal neuron cultures were transfected (three independent experiments) with either miR-101 inhibitor or the miR inhibitor negative control. Protein extracts from hippocampal neurons were analyzed 96 h after transfection by immunoblotting using an antibody specifically recognizing fulllength APP (4G8) or an antibody recognizing GAPDH. Lanes 1 and 4, 2 and 5, 3 and 6 correspond to the first, second, and third transfection experiment, respectively. B, intensities of the bands for each miR inhibitor were quantified by densitometry and the results obtained with the 4G8 antibody were normalized to the anti-GAPDH signal and expressed as arbitrary units of optical density (OD). Results are expressed as means Ϯ S.D. *, p Ͻ 0.05 (t test). tion between miR-101 and its putative target (Fig. 2D). A similar analysis was performed using rat hippocampal tissues from animals from postnatal day 8 up to 6 months of age. The inverse correlation between levels of miR-101 and APP was confirmed (Fig. 3, A-C).
Functional Analysis of miR-101 REs within the APP 3Ј-UTR-To assess whether APP is a target of miR-101, PC12 cells were co-transfected either with a reporter construct containing 1100 base pairs of the human APP 3Ј-UTR downstream of the luciferase open reading frame (Fig. 4A) or with a control firefly plasmid together with either a synthetic microRNA 101 precursor or a control microRNA. We determined that miR-101 significantly suppressed luciferase expression with respect to control microRNA (Fig. 4B). The function of the two miR-101 REs in the APP 3Ј-UTR was evaluated by site-directed mutagenesis of the nucleotides at positions 4 and 5 of the microRNA 101 REs at 242-248 bp (site 1) and position 531-537 bp (site 2) of the APP 3Ј-UTR (Fig. 4A). When cotransfected with miR-101, the reporter carrying mutations in site 2 was still repressed while both the reporter vector carrying mutations in site 1 and the reporter carrying mutations in both sites were insensitive to miR-101-mediated inhibition. These data suggest that modification of the site 1 RE of the APP 3Ј-UTR is sufficient to block the inhibitory function of miR-101 (Fig. 4B). Thus, miR-101 functionally interacts with the APP 3Ј-UTR.

miR-101 Modulates APP Expression in Hippocampal Neurons-To
address whether miR-101 may regulate endogenous APP expression in cultured hippocampal neurons, we performed loss of function experiments by transfecting cells with specific microRNA hairpin inhibitors (Dharmacon). These molecules contain double-stranded flanking regions, around the reverse complement microRNA sequence, which increases the inhibitor function (41). We found that inhibition of endogenous miR-101 significantly up-regulated APP protein levels (Fig. 5, A and B), indicating that hippocampal APP expression may be regulated by miR-101.
The effect of overexpressing miR-101 in differentiated hippocampal neurons was evaluated using a bicistronic lentiviral vector expressing enhanced green fluorescent protein (EGFP) under the control of the CMV promoter, in which the miR-101 sequence was inserted in an artificial hairpin structure under the control of the U6 promoter.
Immunofluorescence and Western blot analysis demonstrated that increases in miR-101 level strongly down-regulated APP protein levels (Fig. 6, B-D). A corresponding slight downregulation of APP mRNA was observed (Fig. 6E), suggesting that although miR-101 may act primarily by reducing APP translation it may also affect APP mRNA stability.
Because the proteolytic products of APP and A␤ deposits are involved in AD, the presence of the soluble N-terminal fragment of APP (sAPP) and A␤ fibrils in conditioned medium from lentiviral transduced hippocampal neurons was determined. A consistent reduction in sAPP and A␤ fibrils was observed in hippocampal neurons overexpressing miR-101 (Fig. 6, F and G). Our findings indicate that miR-101 is a negative regulator of APP expression and suggest that in pathological conditions, this microRNA may protect hippocampal neurons from A␤ load.
miR-101 Contributes to the Regulation of APP in Response to IL-l␤-Although the role of IL-1␤ in regulating APP has been previously analyzed in various cellular types, only a few studies have been performed on primary neuronal cell cultures (43). Hippocampal neurons cultured for 7 DIV were treated with . c, nuclei were labeled with Hoechst 33258 (0.5 mg/ml) (blue). d, merged image showing a decrease in APP staining intensity in hippocampal neurons overexpressing miR-101 relative to control neurons (arrow). C, representative immunoblot analyzing protein extracts from lentivirus infected rat hippocampal neurons at 96 h post-infection using an antibody specifically recognizing full-length APP (4G8) or an antibody recognizing GAPDH. D, intensity of the bands from three independent experiments was quantified relative to pLB-scramble levels, as described in the legend to Fig. 5. Results are expressed as means Ϯ S.D. **, p Ͻ 0.01 (t test). E, qRT-PCR for APP mRNA using hippocampal cell total RNA. Expression relative to pLB-scramble transduced cells with the means from three independent experiments are shown. *, p Ͻ 0.05. F, culture medium conditioned from lentiviral transduced hippocampal neurons was collected from ten different 35-mm culture dishes. Samples were normalized by protein content of the cellular extracts and analyzed by Western blotting using the 22C11 antibody, which recognizes the secreted form of APP. Cultures transfected with pLB-101 possessed sAPP levels that were 0.61 of pLB-scramble control levels. G, from the same conditioned medium A␤ fibrils were isolated by ultracentrifugation and separated on 10 -16% Tris-Tricine gels, electroblotted, and quantified using the 4G8 antibody (4G8 hybridization signal was confirmed with the 6E10 antibody, data not shown). Synthetic A␤1-42 was used as a positive control.
IL-1␤ for up to 8 h and the expression of APP both at the mRNA and protein level was analyzed. While the level of APP mRNA was unchanged after treatment with IL-l␤ (Fig. 7A), the APP protein level was transiently up-regulated, and the amount of APP was increased at 4 and 6 h, but returned to the pretreatment level after 8 h of IL-1␤ stimulation (Fig. 7B). This result suggests that, in hippocampal neurons, the control of APP expression induced by IL-l␤ treatment is exerted primarily at the translational level. To understand whether miR-101 plays a role in the translational regulation of APP induced by IL-l␤, the expression of miR-101 after IL-l␤ stimulation was analyzed. Quantitative RT-PCR analysis showed that mature miR-101 expression was significantly induced by 4 h and reached a peak at 6 h, which was maintained up to 8 h after the initiation of IL-l␤ treatment (Fig. 7C). As shown in Fig. 7D, up-regulation of miR-101 after IL-l␤ stimulation exhibited kinetics distinct from up-regulation of APP. The fact that APP and its negative regulator miR-101 are both transiently up-regulated after IL-1␤ treatment suggests that APP expression cannot be explained solely by the action of miR-101 and that additional regulatory pathways must be involved. In order to evaluate whether IL-1␤-induced miR-101 up-regulation may contribute to the reduction in APP expression in the late phase of IL-1␤ treatment, miR-101 loss of function experiments were performed. After 8 h of IL-1␤ treatment, APP was still up-regulated in neurons in which miR-101 was inhibited (Fig. 7E), suggesting that miR-101 up-regulation may contribute to modulate APP levels under stress conditions.

DISCUSSION
APP is one of the genes potentially regulated by the microRNA pathway in hippocampal neurons, since APP protein levels were up-regulated in neurons in which Ago2 was silenced. Reducing Ago2 levels did not significantly affect APP mRNA levels, suggesting that APP up-regulation is not the consequence of an indirect effect on APP transcription.
The experiments presented here suggest that APP mRNA may be loaded into the RISC complex and may be directly regulated by specific microRNAs. Among microRNAs potentially targeting the APP 3Ј-UTR, miR-101 is a microRNA with two putative REs within the APP 3Ј-UTR and is also expressed in adult hippocampal tissue. Expression of miR-101 and APP both in embryonic primary hippocampal cell cultures and in postnatal rat hippocampal tissues further support the hypothesis that miR-101 is a repressor of hippocampal APP expression. Indeed, using a luciferase assay, we demonstrated that miR-101 actively represses a reporter containing the APP 3Ј-UTR. In addition, using site directed mutagenesis, a functional interaction between miR-101 and one of two microRNA REs within the APP 3Ј-UTR was identified. A number of cis regulatory elements which stabilize or destabilize APP mRNA have been described within the APP 3Ј-UTR (44,45). Moreover, a recent report demonstrated that a RE for miRNAs belonging to the miR-20 family down-regulates APP expression A, hippocampal neurons cultured for 7 DIV were treated with IL-1␤ (10 ng/ml) for 4, 6, or 8 h. A, APP mRNA expression levels were measured using qRT-PCR of total RNA. Expression is presented relative to untreated neurons. Means from three independent experiments are shown. B, protein extracts from hippocampal neurons (three independent experiments) were analyzed by immunoblotting using an antibody specifically recognizing full-length APP (4G8) or an antibody-recognizing GAPDH. The intensity of the bands was analyzed by densitometry, and the results obtained with the 4G8 antibody were normalized to the anti-GAPDH signal and plotted as relative to untreated neurons. Results are expressed as means Ϯ S.D. *, p Ͻ 0.05; **, p Ͻ 0.01 (t test). C, mature miR-101 expression was analyzed by real-time RT-PCR of total RNA. ⌬Ct values were normalized to U6. Expression is presented relative to untreated hippocampal neurons. Means from three independent experiments are shown. *, p Ͻ 0.05; **, p Ͻ 0.01 (t test). D, time course of the increase in APP protein levels and mature miR-101 during IL-1␤ treatment is shown. E, expression of APP in IL-1␤-treated hippocampal neurons after miR-101 inhibition. 3 DIV hippocampal neurons transfected with microRNA hairpin inhibitors were grown until 7 DIV and treated with IL-1␤ for 8 h. The histogram summarizes the results of immunoblotting APP before and after 8 h of IL-1␤ treatment. After miR-101 inhibition, 8 h of IL-1␤ treatment resulted in a 1.33-fold increase in APP relative to untreated neurons. Data are the mean of six experiments Ϯ S.D. *, p Ͻ 0.05. (32). Taken together, these findings indicate that the APP 3Ј-UTR is the target of multiple regulatory pathways, which may constitute an integrated network. Further work will be required to define the functional interactions between the different regulatory elements.
Gain and loss of function experiments identified miR-101 as a negative regulator of APP in rat hippocampal neurons. MiR-101 overexpression reduced A␤ load and Cox-2 levels in neuronal cultures. A␤ assembly into oligomers and fibrils is involved in AD pathogenesis (1) and elevated levels of Cox-2, which contribute to neuroinflammation, are associated with AD (46). Thus, our observations suggest that miR-101 could play a protective role in AD. In fact, the Cox-2 reduction induced by miR-101 overexpression may contribute to the regulation of APP metabolism. Cox-2 is essential for the synthesis of prostaglandin E2 (PGE2) and PGE2 has been shown to induce synthesis of APP mRNA and holoprotein (47) and to stimulate A␤ production (48). Thus, the robust reduction of intracellular APP and of A␤ fibrils in the extracellular space may suggest that miR-101 can act by down-regulating both APP and Cox-2, which also reduces PGE2 levels. Whether other putative miR-101 targets may be associated with APP and A␤ metabolism is an interesting possibility that we will be addressed in future work.
The influence of IL-1␤ on APP expression has been analyzed in various cell types and, in primary neuronal cell cultures, IL-1␤ was able to up-regulate the transcription of APP (43). Here, we demonstrated that treatment of hippocampal neuron cultures with lower IL-1␤ concentrations increased APP through regulation of translation, since the steady-state level of APP mRNA did not increase after IL-1␤ treatment but APP protein levels were transiently up-regulated. In addition, miR-101 may contribute to the reduction in APP expression after prolonged IL-1␤ treatment, suggesting a role for miR-101 in the control of APP expression in response to IL-1␤. However, the overall regulation of APP translation may be the result of multiple mechanisms. For instance, previous findings have suggested that, in human astrocytes (9), IL-1␤ stimulation regulates translation of APP through the 5Ј-UTR of the APP mRNA. Thus, further dissection of the precise role of miR-101 in regulating APP and its potential interaction with other regulatory pathways will be necessary.
Altered expression of microRNAs has been described in AD patients. Two independent miRNA expression profiles (30,49) have shown that miR-101 is down-regulated in the human AD cerebral cortex, suggesting that miR-101 down-regulation may play a role in the development of AD. This hypothesis is in agreement with our results that indicate a direct role of miR-101 in controlling APP translation and A␤ fibril accumulation. Additional studies in AD murine models will be necessary to define the significance of miR-101 in the development and progression of AD pathology.