Cognitive Decline and Modulation of Alzheimer’s Disease-Related Genes After Inhibition of MicroRNA-101 in Mouse Hippocampal Neurons

MicroRNAs have emerged as regulators of brain development and function. Reduction of miR-101 expression has been reported in rodent hippocampus during ageing, in the brain of Alzheimer’s disease (AD) patients and in AD animal models. In this study, we investigated the behavioral and molecular consequences of inhibition of endogenous miR-101 in 4–5-month-old C57BL/6J mice, infused with lentiviral particles expressing a miR-101 sponge (pLSyn-miR-101 sponge) in the CA1 field of the hippocampus. The sponge-infected mouse model showed cognitive impairment. The pLSyn-miR-101 sponge-infected mice were unable to discriminate either a novel object location or a novel object as assessed by object place recognition (OPR) and novel object recognition (NOR) tasks, respectively. Moreover, the sponge-infected mice evaluated for contextual memory in inhibitory avoidance task showed shorter retention latency compared to control pLSyn mice. These cognitive impairment features were associated with increased hippocampal expression of relevant miR-101 target genes, amyloid precursor protein (APP), RanBP9 and Rab5 and overproduction of amyloid beta (Aβ) 42 levels, the more toxic species of Aβ peptide. Notably, phosphorylation-dependent AMP-activated protein kinase (AMPK) hyperactivation is associated with AD pathology and age-dependent memory decline, and we found AMPK hyperphosphorylation in the hippocampus of pLSyn-miR-101 sponge mice. This study demonstrates that mimicking age-associated loss of miR-101 in hippocampal neurons induces cognitive decline and modulation of AD-related genes in mice.


Introduction
Memory impairment, typically associated with synaptic alterations in both the hippocampus and neocortex, is an early sign of AD. Cognitive decline in AD is also associated to Aβ accumulation and in the later stage, to Tau protein alterations and depositions [1]. In addition to fully penetrant mutations in APP, Presenilin 1 and Presenilin 2 causing the familial forms C. Barbato and G. Giacovazzo contributed equally to this work.
Electronic supplementary material The online version of this article (https://doi.org/10.1007/s12035-020-01957-8) contains supplementary material, which is available to authorized users. of the disease, ageing is the main risk factor for AD [2]. The hippocampus, a brain area critical for learning and memory, is among the most vulnerable region in ageing [3]. Cellular changes in neuronal populations, including those occurring in ageing, may lead to increased disease susceptibility [2].
MicroRNA (miRNA)-mediated regulation has been implicated both in learning/memory processes and AD [4]. The miRNAs are small noncoding RNA molecules (20)(21)(22) nucleotides in length), which repress gene target by base-pairing to mRNA [5]. Each miRNA may regulate multiple mRNA targets. In brain tissues, miRNAs may tune synaptic development and plasticity and may modulate brain functions such as memory formation and behavioral performance [6].
MiR-101 is expressed in the brain and also in other tissues, and its modulation has been associated to several biological functions. Noteworthy, the reduction of miR-101 expression has been reported in the cerebral cortex of AD brain [7,8], and miR-101 decrease in the cerebrospinal fluid of AD patients has been correlated with an increase in plaque density [9]. In mice, it has been shown that miR-101 is expressed in CA1 pyramidal neurons as well as in other brain cell types [10]. Decrease of the expression levels of miR-101 in the hippocampus has been reported in mice and rats during ageing [11][12][13]. Furthermore, miR-101 is downregulated in the hippocampus of very old and young mouse models of AD compared with their age-matched normal controls [11] and in the cerebral cortex of APP/PS1 mice [14].
Previously, we demonstrated that miR-101 may directly regulate AD-related genes APP and RanBP9 in rodent hippocampal neurons both in vitro and in vivo [15,16] and that miR-101-mediated post-transcriptional regulation of RanBP9 may modulate the amyloidogenic processing of APP [16]. Consistently, miR-101-mediated regulation of APP expression has also been demonstrated in human cells [17]. More recently, it has been reported that administration of miR-101b can promote an improvement of hippocampaldependent memory deficits in 12-month-old 3xTg-AD mice thus suggesting a protective role for this microRNA during AD progression [18].
Here, we investigated the impact of endogenous hippocampal miR-101 inhibition on cognitive decline and the associated molecular changes by sponging this miRNA in CA1 neurons of wild-type C57BL/6J male mice.

Preparation of Lentiviral Particles
pLSyn and pLSyn-miR-101 sponge lentiviral vectors as well as the preparation of the G glycoprotein vesicular stomatitis virus-pseudotyped lentiviral particles have been previously described [16]. Briefly, miR-101 sponge is a transcript under the control of the synapsin promoter and carries four tandem copies of a sequence complementary to the miR-101 located downstream of the enhanced green fluorescent protein (EGFP) open reading frame [16]. The titers of the viral vectors used in this study were in the range of 1-2 × 10 9 transducing units (TU)/mL.

Cell Culture
Primary hippocampal neurons were prepared from day 17 to 18 embryos obtained from timed-pregnant Wistar rats (Charles River) as previously reported [15]. Half of the medium was changed every 3-4 days. The neurons were transduced at 7 days in vitro with 2 × 10 6 TU/mL of the pLSyn-miR-101 sponge vector, or with a control pLSyn vector expressing only EGFP. Protein lysates were collected at 7 days post-transduction and used in Western blotting experiments, as described below.

Animals
C57BL/6J male mice (provided by the European Mouse Mutant Archive (EMMA) facility in Monterotondo, Rome, Italy) were housed in standard 12-h light/dark cycle and used at the age of 3-4 months. All experiments were performed in accordance with the Italian Ministry of Health and European Directive 2010/63/EU. Mice were bilaterally microinfused into the CA1 field of the hippocampus with lentiviral particles derived from the control pLSyn vector or the pLSyn-miR-101 sponge vector.
After surgery mice were intraperitoneally administered with sterile saline (NaCl 0.9%; volume 10 mL/kg) to prevent dehydration and monitored post-operatively until they showed undisturbed free motor activity. Four weeks later, animals were assessed in behavioral tasks, as described below.

Cognitive Assessment
To investigate AD-like cognitive impairment, animals that underwent intrahippocampal miR-101 inhibition were assessed for their performances in object place recognition (OPR) task and novel object recognition (NOR) task. The OPR task was used to evaluate hippocampus-dependent spatial learning, while the NOR the hippocampal and nonhippocampal-dependent short-term recognition memory. Later, to evaluate contextual associative memory mice were also assessed in a delay fear conditioning (FC) task. Both OPR and NOR were assessed by inserting the animals into a brightly-lit circular open field (i.e., arena) of about 60 cm diameter closed by a wall of 20 cm high. The apparatus used is made of gray plastic with the floor of the open field divided into several sectors by black lines. Other detailed features of the apparatus were previously reported [19][20][21]. During both the OPR and NOR performances, the arena was placed into a soundproof room. Attached to the wall of the apparatus, there was a striped pattern (20 cm wide and 10 cm high), as a spatial cue reference. A red light (80 W) illuminated the arena. Performances were continuously recorded by a video camera placed just above the open field. Five objects were simultaneously present in the open field: a chromium-plated parallelepiped (7 × 4 × 4 cm); a plastic cone on a transparent cylindrical base (height 6 cm, diameter 8 cm); a small ladder-like item made of gray plastic material (height 16 cm, width 5 cm, number of steps 10) inserted on a cylindrical base (height 2 cm, diameter 7 cm); a plastic cylinder (height 10 cm, diameter 5 cm) on a transparent Plexiglas base with a nut (height 2 cm) fixed on the top; and a transparent plastic spool (height 12 cm, diameter of the top and the base 5 cm). The initial arrangement consisted of placing a square-shaped object (plastic cone) in a central position. The cognitive testing procedure was detailed previously [19,21]. For both OPR and NOR, the experimenter touched and manipulated all objects before each session.

Object Place Recognition
On the test day, mice were placed into the empty arena for a 20-min session, to get acquainted with the apparatus and to determine the baseline level of locomotor activity. After a 5min interval, each animal was placed back in the open field (with objects) for five successive 6-min sessions, separated by a 5-min interval, during which mice were returned to their cages. During sessions 2-4 (S2-S4), the objects were placed in a square-like configuration, with a central object (cone, object B). In S5 (OPR) session, the configuration was changed by displacing two objects: the cone (B) replaced the cylinder (D), which was itself displaced at the periphery of the open field (between ladder (C) and parallelepiped (A)), so that the initial square arrangement was changed to a new spatial arrangement, thus challenging the animals to perform a novel OPR.

Novel Object Recognition
In the subsequent 6-min last session (S6), mice were assessed for their performance in NOR. An additional unfamiliar object, consisting of an iron squares (10 × 10 cm) forming an angle of 90°(corner-shaped object), was used to test the ability to perform the NOR task. Hence, one of the familiar nondisplaced objects (NDOs; spool, object E) was replaced by a new object (corner, object F) in the same location.

Passive (Inhibitory) Avoidance Task
The ability to retrieve a contextual fear memory was investigated by the use of an inhibitory avoidance task. Here, the inhibitory avoidance conditioning was studied by the onetrial learning step-through procedure. Briefly, the stepthrough inhibitory avoidance apparatus consisted in a stainless steel grid-floor box (57 × 27 × 30 cm) composed of acrylic walls and two compartments, a bright (white) illuminated (350 lx lamp) partition and a dark (black) partition separated in the middle by a guillotine door. The apparatus is also equipped with a control unit to deliver a scrambler shock (Ugo Basile Biological Apparatus, Comerio, Varese, Italy). On training day (day 1), mice were left in the experimental room for 1 h before the start of the experimental procedure, and then, each mouse was habituated for 10 min to the apparatus. The training session was carried out the day after (24 h later) by placing each mouse in the bright compartment. After 5 s, the guillotine door was opened and each animal was allowed to step-through with the four paws into the dark compartment where a foot shock (0.5 mA, 3 s duration) was immediately delivered. Then, the animal was taken out and returned to the home cage. Twenty-four hours after the contextual fear conditioning, the procedure was repeated to evaluate the ability to retrieve the contextual fear memory. Thus, each mouse was placed again in the bright side, the door opened after 5 s and the latency to re-access to the dark compartment recorded by considering a maximum time (cut-off) of 300 s, during which no electric foot-shock was delivered.

Statistical Analysis
Statistical analysis of the behavioral data was performed only in mice that bilaterally expressed the GFP in the hippocampus as assessed by Western blotting. Data (mean ± SEM) obtained from the assessment of both OPR and NOR performances were analyzed by two-way analysis of variance (ANOVA) with treatment (pL-Syn vs pL-Syn miR-101 sponge) as between-group factors and objects (either NP vs FP or NO vs FO) as within-subject factors. ANOVA analysis was followed by post hoc multiple comparisons (Tukey HSD) between pL-Syn and pL-Syn miR-101 groups and between place (novel place vs familiar place) and object (novel object vs familiar object) conditions. Locomotor activity (total distance traveled, cm) as well as performances (latency, s) of memory retention in the step-through inhibitory avoidance were analyzed by comparing pL-Syn and pL-Syn miR-101 groups by unpaired Student's t test. All the analyses were performed by using GraphPad Prism version 8.0. (GraphPad Software, San Diego, California USA). P < 0.05 was considered to indicate a significant statistical difference. Moreover, to improve transparency of data representation, we did not utilize the commonly exploited bar graphs instead we used box and dot plots, including median (and not mean), upper and lower variations and all the actual data-points, as also suggested elsewhere (e.g., https://thenode.biologists.com/ leaving-bar-five-steps/).

Protein Extraction and Western Blot Analysis
Hippocampal tissues, harvested and flash-frozen in dry ice, or cultured cells were homogenized in buffer (1% Triton, 0.25% sodium dodecyl sulphate (SDS), 1% sodium deoxycholate, 2 mM EDTA and 1 mM dithiothreitol) supplemented with a protease and phosphatase inhibitor mixture (Sigma-Aldrich) to yield total protein extracts. The standard 2× Laemmli loading buffer contained 4% SDS, 10% β-mercaptoethanol, 20% glycerol and 0.04% bromophenol blue in 125 mM Tris-HCl, pH 6.8. Equal amounts of total protein extract were fractionated by electrophoresis in an 8% or 10% SDS-polyacrylamide gel and then transferred to a nitrocellulose or PVDF membrane (Hybond-ECL, GE Healthcare; Millipore). Membranes were incubated with the indicated primary antibody overnight at 4°C. Incubation with a secondary peroxidase-coupled antimouse or anti-rabbit antibody (GE Healthcare) was performed at room temperature for 1 h. Immunoreactivity was determined by using an enhanced chemiluminescence detection kit (Millipore). The following primary antibodies and dilutions were used: rabbit monoclonal anti-phospho-AMPKα (Thr172) (D79.5E Cell Signaling, RRID: AB_2169396, 1:1000), Rabbit pAb to AMPK alpha 1 Abcam AB3759, RRID: AB_304054, 1:1000), mouse monoclonal anti-APP (4G8, previously Covance SIG-39220, Biolegend RRID: . Autoradiography films for APP, Rab5 and AMPKα phosphorylation (Thr172) and AMPKα 1 were scanned and band intensities were calculated using the Quantity One software and normalized to GAPDH or α-Actin, or tubulin signal. A Chemidoc XRS Bio-rad with a chemiluminescent camera and Image Lab software 4.0 was used for acquisition and quantification of RanBP9 signal normalized to α-Actin signal.

Lysis of Hippocampal Tissues and ELISA
Four hippocampi obtained from different mice expressing the miR-101 sponge and 3 hippocampi from different pLSyn control mice were flash-frozen, extracted with 70% formic acid and sonicated. Insoluble material was removed by centrifugation to 40,000×g for 20 min. The supernatants were neutralized and diluted 1:20 with Tris base 2 M. Samples were analyzed with human/rat β amyloid (42) ELISA kit (Wako Human/Rat (Mouse) β-Amyloid (42) ELISA High-Sensitive Kit; Catalog Number: 292-64501) [22]. ELISA was performed according to the manufacturer's protocol.

Immunohistochemistry
Mice were intracardially perfused with 4% paraformaldehyde (PFA). Brain were removed and post-fixed overnight in the same solution and then kept in 30% sucrose for 48 h. Brain were coronally sectioned on a vibratome Leica. Free floating sections of 40 μm encompassing the hippocampus were permeabilized in blocking solution (0.3% Triton × 100 and 10% normal goat serum in PBS) for 30 min and incubated overnight with mouse monoclonal antibody anti-NeuN (A60 clone Millipore, RRID: AB_10807694) or mouse monoclonal anti-Amyloid Beta (Aβ) 42 (clone 12F4 Biolegend, RRID: AB_2564683, 1:50). Sections were then incubated with goat anti-mouse IgG Alexa 555 for 2 h. Nuclei were stained with 4,6-diamidine-2-phenylindole dihydrochloride (DAPI; Sigma-Aldrich; 50 mg/ml in PBS for 15 min at RT). Sections were mounted with fluoromount (Sigma) and analyzed at confocal microscopy TCS SP5 microscope (Leica Microsystem). Three-four brain slices, encompassing CA1 hippocampal neurons, GFP-positive and stained with the 12F4 antibody were imaged. A total of 3 miR-101 sponge-infected mice and 3 control mice were studied. Zstack images were captured at 1-μm intervals with a 40× objective and a pinhole 1.0 Airy unit. For imaging of sections stained with the 12F4 antibody zoom of 2 or 3.6 were used.

Cognitive Impairment in Mice Expressing the miR-101 Sponge in CA1 Hippocampal Neurons
To investigate whether mimicking age-and AD-associated miR-101 decrease in hippocampal neurons may lead to impairment of hippocampus-dependent memory in adult mice, we used the pLSyn-miR-101 sponge lentiviral vector and C57BL/6J male mice. This mouse model was chosen because it has been previously reported that miR101 is downregulated in the hippocampus of aged wild-type male C57BL/6J mice [11,13] and of transgenic AD mice with a C57BL/6J background [11]. MiRNA Fig. 1 Schematic representation of the experimental design. a Infusion of lentiviral pLSyn or pLSyn-miR-101 sponge particles in the CA1 region of the hippocampus of 12-16-week-old mice. b Four weeks later, the mice hippocampi were analyzed. Validation of CA1 injection in a pilot experiment in mice infused with pLSyn; EGFP-positive, CA1 hippocampal neurons, stained also with a neuronal nuclei (NeuN) specific antibody (red). c Simplified illustration of the different consecutive sessions used to implement the OPR and NOR tasks. d Simplified illustration of the step-through inhibitory avoidance task to assess long-term contextual memory retrieval. e EGFP expression by Western blotting for the selection of mice injected bilaterally in the hippocampi (red square) (L, left and R, right). f Statistical analysis of the behavioral performances and expression of miR-101 target genes sponge transcripts act as competitive inhibitors by sequestering miRNA of interest [23]. The pLSyn-miR-101 sponge, as previously reported, has neuronal selectivity and efficiently inhibits the miR-101 action on its targets in rodent hippocampal neurons both in vitro and in vitro [16].
Next, lentiviral particles were infused in male mice to perform behavioral experiments (Fig. 1c, d) and only mice positive for bilateral EGFP expression in the hippocampus were considered for statistical and biochemical analysis (Fig. 1e, f; Supplementary Fig. 1). Regarding the efficiency of the intrahippocampal infusions, 50-60% of the mice r es u l t e d E G F P p o s i t i v e b o t h in l e f t a n d r i gh t hippocampus.
Episodic memory impairment and alteration of spatial orientation are associated to the early phase of AD; as the disease progresses, non-spatial working memory Fig. 2 miR-101 inhibition in CA1 hippocampal neurons impairs different aspects of hippocampus-dependent learning and memory. a The ability of mice to perform the OPR task; it depicts the box and whisker plot, with individual data, of the time (s) spent in contact with novel place (NP) objects or with familiar place (FP) objects in session 5, minus the time spent in contact with the same non-displaced objects in the previous session 4 (last session of habituation). b The ability to perform the NOR task; it depicts the box and whisker plot, with individual data, of the time (s) in contact with the novel object (NO) or with the familiar objects (FO) in session 6, minus the time spent in contact with the same non-replaced objects in the previous session 5 (session of novel objects placement). **P < 0.001 NP pLSyn vs FP pLSyn;°°P < 0.001 NP pLSyn vs NP pLSyn-miR-101. c The box and whisker plot, with individual data, of the motor activity expressed by mice during sessions from 2 to 6. d The step-through latency of inhibitory avoidance conditioning test and the deficit of memory retrieval in pLSyn-miR-101infused mice. It depicts the box and whisker plot, with individual data, of the step-through latency (s). ***P < 0.0001 pLSyn vs pLSyn-miR-101. For all box and whisker plots, middle lines show median value whereas box hinges the lower (first) and the upper (third) quartiles, and whiskers minimum and maximum range of time of exploration becomes poorer. Thus, we asked whether inhibition of miR-101 in CA1 neurons might affect object exploration and response to spatial and nonspatial cues according to a previously described procedure [19][20][21]. During the retention trial of the spatial novelty task (5 min after the training trials), C57BL/6J mice expressing the miR-101 sponge in the hippocampus were unable to distinguish either displaced object or novel object respect to familiar object (Fig. 2a, b). Importantly, no differences in the locomotor activity of the pLSyn-miR-101 sponge infused mice with respect to mice infused with the pLSyn control vector were detected (Fig. 2c). The same groups of mice were also tested in the inhibitory avoidance task, and 24 h after training, shorter retention latency was observed in the pLSyn-miR-101 sponge mice with respect to control pLSyn mice (Fig. 2d). We could exclude experimental bias due to previous testing, because the training of a new group of naïve mice in the inhibitory avoidance task confirmed a memory impairment associated to miR-101 sponge expression (data not shown). Thus, miR-101 inhibition in CA1 hippocampal neurons impaired both short-and long-term memory processes.

Increased Protein Levels of AD-Related miR-101 Target Genes in the pLSyn-miR-101-Injected Mice: Correlation with Cognitive Impairment
Having established that the expression of the miR-101 sponge in CA1 hippocampal neurons affects behavioral performances in adult mice, the expression of selected miR-101 targets in hippocampal tissues was investigated in mice engaged in the behavioral tasks. Among several validated miR-101 target genes, APP and RanBP9 genes [15][16][17] were selected because they were known to be associated to AD. In particular, APP and RanBP9 transgenic mice have been reported to be impaired in spatial learning and memory skills compared to wild-type control mice [24][25][26][27]. Rab5, a miR-101 target gene [28], was also selected because its upregulation has been described within human hippocampus in mild cognitive impairment (MCI) and AD [29]. Western blotting experiments demonstrated that endogenous APP, RanBP9 and Rab5 protein levels were significantly higher in extracts of hippocampi derived from pLSyn-miR-101 sponge mice compared to pLSyn mice ( Fig. 3; Supplementary Fig. 1).
Excessive increase of Aβ peptide is one of the main pathogenic event linked to AD and relative cognitive impairments Fig. 3 Increase of protein expression levels of miR-101 target genes in the hippocampus of mice with memory impairment. a Representative immunoblots of App, detected with both 4G8 mAb (reactive to 17-24 aa of Aβ region of APP) and A8717 antibody (reactive to the APP Cterminal region), of RanBP9 and Rab5 in hippocampi of pLSyn-miR-101 sponge and pLSyn mice. Actin was used for normalization. b Means ± SEM of APP, RanBP9, Rab5 protein levels obtained from three independent experiments are presented relative to control pLSyn samples (*p < 0.05, Student's t test) [30][31][32], although picomolar concentration of Aβ peptides may drive protective effects on memory and synaptic plasticity processes [33][34][35]. As Aβ levels can be modulated by genes regulated by miR-101 [15,16] and APP and RanBP9 proteins were indeed upregulated in the pLSyn-miR-101 sponge infused mice with respect to pLSyn control mice, we asked whether miR-101 inhibition in the hippocampus could modulate the levels of Aβ42 peptides.
First, soluble/insoluble Aβ42 peptides were measured in the hippocampus of mice expressing the miR-101 sponge by using a sandwich ELISA. We found that miR-101 inhibition may modulate Aβ accumulation; in fact, significant higher levels of Aβ42 peptides were detected in the hippocampus extracts of mice infused with the miR-101 sponge (Fig. 4a). Moreover, we observed Aβ intraneuronal accumulation by Aβ42 staining in the CA1 region of the hippocampus of pLSyn-miR-101 sponge infused and cognitively impaired mice (Fig. 4b). Aβ signals were associated to fields with high EGFP expression (Fig. 4b) but also with lower ones (Supplementary Fig. 2).
Overall, these results demonstrate the efficacy of the miRNA-101 sponge on the regulation of target genes in mice with cognitive impairment. These findings raise the possibility that miR-101 by modulating AD-related genes might contribute to the cognitive deficits.
Increased AMPK Phosphorylation at Thr 172 in the Hippocampus of pLSyn-miR-101 Sponge Mice Dysregulation of multiple mechanisms has been linked to ADlike pathology and in particular to synaptic failure as well as to hippocampal-dependent cognitive impairment. MiR-101 inhibition may affect several target genes beyond those previously mentioned including AMPK α [18]. Over-activation of AMPK may be detrimental for synaptic plasticity and memory processes in ageing and AD [36][37][38][39]. Furthermore, the High level of Aβ42 staining in pLSyn-miR-101 sponge neurons (lower panels) is indicated by arrows. For each mouse, 3-4 brain slices encompassing CA1 hippocampal neurons were imaged. A total of three miR-101 sponge-infected mice and three control mice were studied. Scale bar 50 μm, 40× magnification, zoom 2 (left panels) and 25 μm, 40× magnification, zoom 3.6 (right panels). Right panels represent 1.8× magnification of the corresponding dot-lines insets expression of a constitutive active AMPK construct in rat CA1 hippocampal neurons has been reported to impair contextual fear conditioning memory [40]. In order to identify molecular correlates of miR-101 sponge-induced cognitive impairment, the levels of AMPK activation were analyzed in the hippocampi of pLSyn-miR-101 sponge mice with respect to control mice (Fig. 5, Supplementary Fig.  1). We found that phosphorylation of the AMPK αsubunit at Thr 172 was significantly increased in the experimental mice compared with the pLSyn control mice (Fig. 5a), while the expression levels of total AMPK α1 were similar in pLSyn-miR101 sponge and control tissue lysates. Experiments in cultured primary hippocampal neurons confirmed that the miR-101 sponge vector could induce AMPK activation while it does not affect total AMPK α1 protein expression levels (Fig. 5b). These findings suggest that AMPK deregulation during memory encoding may contribute to the cognitive deficits observed in the pLSyn-miR-101 sponge infused mice.

Discussion
Inhibition of miR-101 post-transcriptional regulation in CA1 hippocampal neurons of 4-5-month-old C57BL/6J male mice leads to a cognitive decline resembling traits observed in AD transgenic mouse models. Detrimental effects were recorded in three tasks for cognitive evaluation. In particular, mice showed a decline in hippocampal-dependent spatial learning, short-term recognition memory and hippocampal-dependent contextual fear memory. The decline of hippocampal-dependent spatial learning has been reported both in aged mice and AD mouse models [41,42]. Conflicting results have been reported for short-term object recognition memory deficits likely due to the different mouse models used [43][44][45]. Cognitive deficits in contextual fear memory have been shown in several AD transgenic mouse models also at early age [42].
The spatial object location learning is strictly dependent from the hippocampus while the object recognition memory is dependent from several brain regions. Therefore, we can hypothesize that the detrimental effects on NOR ability can be due to alterations in the functionality of other brain regions implicated in NOR task affected indirectly by the miR-101 sponge expressed in the CA1 hippocampal neurons.
The dysregulation of miR-101 has been associated to multiple processes in rodent brain [46,47] including its effect in early hippocampal development in wild-type mice [10] and in the hippocampus of an AD transgenic mouse model [18]. In this study, we demonstrate that miR-101 loss of function in CA1 hippocampal neurons of adult male wild-type mice may affect learning and memory processes. The neuronal selectivity of the miR-101 sponge and its spatial specificity achieved by stereotaxic injection of lentiviral particles allowed us to associate the behavioral impairment to the blockade of miR-101 in CA1 hippocampal neurons. At this stage, we cannot exclude the possibility that miR-101 blockade in CA1 neurons can also indirectly impact the functions of other brain cells and circuitry establishing direct and indirect neural communication with the hippocampus.
MiRNAs target multiple genes that may be different depending on the cell type and stage of development. MiR-101 sponge-induced cognitive impairment here reported in adult mice was correlated to an increase of protein expression of the APP and RanBP9 target genes. We can speculate that such genes are contributing to the observed phenotype because their expression has been shown to be relevant in several transgenic mouse models with learning and memory deficits [25][26][27]48]. We also observed an increase of Aβ42 peptide levels in the miR-101 sponged hippocampi with an intraneuronal accumulation of Aβ42 in CA1 sponged neurons. This finding is of interest because in AD models and Aβ transgenic mice, intraneuronal Aβ accumulation has been shown to trigger cognitive deficits [31,49,50].
Finally, the AMPK hyper-phosphorylation observed in the pLSyn-miR-101 sponge mice is in agreement with the overactivation of AMPK reported in AD patients, AD mouse models [36,38] and in the brain of aged mice [39], while in our experimental condition, we did not observe modulation of the AMPKα protein reported in other studies [18]. This finding raises the possibility that miR101 sponge is indirectly impacting the phosphorylation status of AMPK through other miR101 targets.
Overall, our results show that miR-101 function in adult CA1 hippocampal neurons is required for the normal cognitive processing of cognitive tasks in adult male mice. The target genes investigated suggest that regulation of miR-101 expression can play an important role in the modulation of biochemical pathways known to be involved in the course of AD pathology and ageing. In this miRNA sponge mouse model, a deep investigation of miR-101 target genes and related pathways could be instrumental to dissect the molecular mechanisms involved in hippocampal-dependent cognitive decline after miR-101 downregulation that may occur during ageing and AD.
Funding Information This work was supported by the Consiglio Nazionale delle Ricerche, Aging Programme DSB.AD009.001.004 IBCN Workpackage 1.9 "Study of molecular and cellular mechanisms involved in the pathogenesis and progression of Alzheimer's Disease" to Francesca Ruberti & Christian Barbato. The Italian Ministry of Health (Ricerca Corrente) also supported the work.