MiR-188-5p Affects the Progression of Alzheimer's disease by Negatively Regulating Neuritin Expression

Background Alzheimer's disease (AD) is a progressive neurodegenerative disease characterized by complex pathobiological characteristics and still lacks accurate biomarkers in early diagnosis. Neuritin (NRN1) is a neurotrophic factor, which has active roles in neuronal plasticity and regeneration. Recent research suggests that Neuritin has been associated with loss of cognitive function, maybe a potential therapeutic target in AD. Methods We have clustered the upregulated miRNAs in AD and predicted the target miRNAs of Neuritin by bioinformatics analysis, found miR-188-5p may involve in the development of AD. Results In the present study, we conrm the association between Neuritin and miR-188-5p expression in mice with AD. Besides, we provide evidence for the changes of cognition capacity, miR-188-5p and Neuritin levels in APP/PS1 mice. Conclusion Our results reveal a previously undened mechanism that miR-188-5p plays a signicant role in the development of AD by inhibiting the expression of target protein Neuritin.


Introduction
Alzheimer's disease (AD) is a progressive neurodegenerative disease characterized by the accumulation of amyloid β plaques and by neuro brillary tangles, with eventual progressive cognitive and memory impairment [1][2][3]. Due to its complex pathobiological characteristics, accurate biomarkers in early diagnostic and effective therapeutic options are still lacking in the detection or treatment of AD [4,5].
Recent research suggests that Neuritin has been associated with loss of cognitive function, maybe a potential therapeutic target in AD [9]. However, the mechanisms of Neuritin in neurodegenerative diseases are not fully understood.
The further development of bioinformatics approaches has broadened the eld of microRNA (miRNA) research, especially in target gene prediction and target site accessibility [10][11][12]. Studies have con rmed that miRNA dysregulation plays an important role in the development of a variety of diseases [13] and maybe involved as biomarkers in the progression of AD [14][15][16].
Here, we clustered the up-regulated miRNAs in AD from the literature and datasets of AD and predicted the target miRNAs of Neuritin by bioinformatics analysis, found the association between miR-188-5p and Neuritin expression in mice model with AD (APP/PS1 mice). Also, we provided evidence for the associations between the cognition (working memory) capacity of APP/PS1 mice and altered miR-188-5p and Neuritin levels. Our results reveal a previously unde ned mechanism that miR-188-5p plays a signi cant role in the development of AD by inhibiting the expression of target protein Neuritin.

Ethical statement
All experiments using animals were performed in strict accordance with the Institutional Animal Care and Use Committee (IACUC) and approved by the Animal Ethical Committee of Hangzhou Normal University (Hangzhou, China).

Animals
APPswe/PS1ΔE9(APP/PS1) double-transgenic AD mice and age-matched B6C3 hybrid genetic wild-type (WT) mice were purchased from The Animal Model Center of Nanjing University (Nanjing, China).Mice were housed in same-sex groups and kept on a regular diet and light/dark cycle in a Speci c Pathogen Free (SPF) laboratory animal environmental facilities, in which food and tap water were available ad libitum.Both APP/PS1 AD mice and WT control mice were divided into three age cohorts 1 4, and 7 months of age , at the onset of the experiment. All mice were performed Fear Conditioning test at the age of 4 and 7 months and euthanized for hippocampus and cortex tissue extraction from the brain at the age of 1, 4, and 7 months.

RNA extraction and quantitative RT-PCR
The total RNA from the mice hippocampus and cortex was extracted by SanPerp Column microRNA Extraction Kit (Sangon, SK8191) and cDNAs were synthesized from RNA templates by RevertAid Premium Reverse Transcriptase (Thermo Scienti cTM, EP0733). The reverse transcription products were ampli ed by SG Fast qPCR Master Mix (2X, Roche, B639271). The neuritin expression from the mice hippocampus and cortex was evaluated by qPCR using FastKing gDNA Dispelling RT SuperMix Kit (TIANGEN, KR170801). The extraction of total RNA and synthesis of cDNAs was performed with TRIzolTM reagent (Invitrogen, 15596026) according to the manufacturer's protocol. The A260/A280 ratio was recorded and the concentration and the purity of RNA were assessed by Nanodrop 2000. The expression levels of miRNA and genes were normalized to U6 snRNA levels and GAPDH, respectively.
The primers were designed by the Primer Premier 5.0 Software and as follows: with the matched secondary antibody with HRP for 2 hours at room temperature. The bands were visualized with an enhanced chemiluminescence system and protein levels of Neuritin were normalized to the bands of β-actin.

Bioinformatics analysis of candidate miRNAs
To obtain miRNA datasets, by searching for miRNA expression pro ling data in AD from the Gene Expression Omnibus database (http://www.ncbi.nlm.nih.gov/geo/), using R language to select criteria samples from AD and NC groups, excluding epigenetic data. Candidate miRNAs were screened from 3 datasets of AD (GSE46579, GSE104514, and GSE16759), by a matrix of miRNA expression levels in the Bioconductor LIMMA R package. The screening threshold of candidate miRNAs was P<0.05, |fold change|>1.
All homo sapiens and mice sequences related to candidate RNAs were retrieved from miRBase (http://www.mirbase.org/), multiple sequence alignment analysis, and phylogenetic tree construction were performed by EMBL-Cluster Omega (http://www.ebi.ac.uk/Tools/msa) and Jalveiw software.

Fearing Conditioning test
Fear Conditioning (FC) test was taken place in 2 chambers (Panlab, LE116, Barcelona, Spain) as described [17][18][19][20]. There was a speaker attached to one of the stainless-steel walls in the chambers, which provided continuous background noise (30 dB) during the test. The stainless-steel grid oor of the chamber could deliver electric foot shocks, from which the mice could not escape. The 4-and 7-monthsold APP/PS1 AD mice (N=5 per age cohort) and age-matched WT mice (N=5 per age cohort) were tested in the whole FC test.
The FC test consisted of 2 phases, the training and the test phase, which took place in 48 consecutive hours. During the training phase, all groups of mice were placed in a standard fear conditioning chamber A (25×25×25 cm) and allowed to acclimate for 120 seconds without stimulation. After acclimation, each mouse underwent 7 cycles of tones (30 seconds per cycle, 80dB) and electric foot shocks (2 seconds duration per cycle, 0.35mA), which stopped with the tones. Each foot shock was separated by a 30-60 seconds time period. The training phase ended with a 30 seconds post-training period, then mice were removed from the chamber and returned to the home cages. During the test phase, the mice underwent fear conditioning testing for both the contextual and cued memory tests, at 3, 24, and 48 hours after the training phase. For the contextual recall test, the mice were placed back in chamber A without any tones and foot shocks for 6 minutes. For the cued recall test, the mice were placed in chamber B (29×25×22 cm, with different walls, oor, smell, and light, compared with chamber A), treated with the same acclimation, tones, and post-training period as the training phase, without foot shocks. For each phase and test, the freezing behavior, which is de ned as the cessation of all movements despite respiratory behavior [21], was recorded. The percentage of the time that the mice displayed freezing behavior was calculated and the differences between groups in time spent freezing were analyzed.
The 293T cells were transfected using miR-188-5p mimics or normal control (NC) miRNA and treated for 24 hours. Then, the cells were collected for later experiments.
Luciferase assay A luciferase activity assay was performed using a Dual-Luciferase Reporter Assay Kit (Promega, USA), according to the manufacturer's instructions.
The primers were designed and applied as follows:

Statistical analysis
Bioinformatics analysis was performed using R software. Statistical analysis was performed using GraphPad 6.0. Differences between two groups were determined using one-way analysis of variance (ANOVA), and multiple comparisons were performed with a t-test posthoc test. All values were presented as the mean ± standard deviation (SD). P<0.05 was considered statistically signi cant.

Results
1 Decrease of Neuritin expression in the hippocampus of APP/PS1 mice.
In the hippocampus and cortex of APP/PS1 mice at 1, 4, and 7 months, the expression of Neuritin mRNA and protein level was evaluated by qPCR analysis and Western blot (Fig. 1). In accordance with the result of qPCR, the expression of Neuritin mRNA was notably lower in the hippocampus of the APP/PS1 mice, compared with the WT mice. And, the Neuritin mRNA expression in the hippocampus of APP/PS1 mice at the lowest level in 4-months old (P<0.01, Fig.1 A). Similarly, the Western blot analysis showed signi cant reductions in the expression level of the Neuritin protein in the hippocampus at 4-and 7-months old APP/PS1 mice, compared with the age-matched WT mice (P<0.05, Fig.1 C, D). However, no group differences were found in the expression level of Neuritin mRNA in cortexes at each age group of mice ( Fig.1 B), nor were differences observed in the expression level of Neuritin protein in the hippocampus of 1-month-old APP/PS1 mice, compared with the WT mice (Fig.3 A, C, D). Revealing that Neuritin level was downregulated in the hippocampus of APP/PS1 mice from the age of 4 months. 2 The selection of candidate miRNAs that may regulate Neuritin in AD by Bioinformatics.
We collected 3 miRNA expression datasets (GSE46579, GSE104514, and GSE16759) from AD patients and AD model mice as well as literature data. R language conductor package was used to normalize the measurement data, standardize the quality and cluster analysis of the 3 datasets. MiRNAs were screened using the cut-off criteria of P-value <0.05 and |log2FC|≥1 by the LIMMA R package. According to the volcano plots, 96 and 51 miRNAs were up-regulated in GSE46579 ( Fig. 2A) and GSE104514 (Fig. 2B), respectively. Candidate miRNAs (miR-188 and miR-184) were obtained by overlapping up-regulated miRNAs (Fig. 2C).
We further searched for the candidate miRNAs that may target the 3'-UTR of Neuritin by 4 bioinformatics algorithms. 481 miRNAs from TargetScan, 92 miRNAs from miRBD, 253 miRNAs from DIANA-microT, and 58 miRNAs from miRWalk were overlapped by VENNY (Fig. 2D). We selected 7 miRNAs target the Neuritin for the candidate miRNAs.
Then, we performed multiple sequence alignment and homology analysis of candidate miRNAs from different species with bioinformatics websites and software (Fig. 2E). The phylogenetic tree showed the homology of miR-188-5p and miR-188-3p in Homo sapiens and mice (Fig. 2F, Green texts), and the global alignment of mmu-miR-188-5p and has-miR-188-5p sequences showed the identity (Fig. 2G).
3 The variations of the candidate miRNAs (miR-188-5p) and Neuritin in the hippocampus of APP/PS1 mice.
We investigated the expression of candidate miRNAs (miR-188-5p) in the hippocampus and cerebral cortexes of the APP/PS1 mice at the ages of 1, 4, and 7 months. MiR-188-5p was highly expressed in the hippocampus in 4 months of APP/PS1 mice (P<0.05, Fig. 3A). In addition, the assays showed that the expression of miR-188-5p was higher in 4 months than 1 and 7 months in the hippocampus of APP/PS1 mice, compared with the WT mice (Fig. 3C). However, the expression level of miR-188-5p in the cerebral cortex was not statistically signi cant, either in the APP/PS1 mice or compared with the WT mice (Fig.  3B). Furthermore, we detected the expression of Neuritin protein from the hippocampus of these mice, Western bolting assay showed the protein level of Neuritin decreased from 4 months in the APP/PS1 mice, compared with WT mice (P<0.05, Fig. 3D, E).

Evaluation of cognition (working memory) capacity in APP/PS1 mice by Fear Conditioning test.
Fear Conditioning test was performed at the age of 4 and 7 months for APP/PS1 and WT mice (Fig. 4A). In the training phase, in which we characterized the exploration of the training chamber (120 seconds) and the presentation of a tone (30 seconds, 7 times), no statistically differences were observed in the freezing times during the response to the administration of foot shock (2 seconds, 0.35mA) between the genotypes and age groups (Fig. 4B), suggesting that APP/PS1 and WT mice showed comparable initial reactivity to experimental stimuli.
In the test phase, mice underwent fear conditioning testing for both the contextual and cued recall. The analysis of the freezing times indicated that APP/PS1 mice at the age of 4 and 7 months spent less time freezing during the contextual fear memory conditioning, showed statistical differences at 3 and 48 hours after the training phase, compared with age-matched WT mice (Fig.4 C, E). On this basis, during the cued fear memory conditioning, a signi cant reduction of freezing times was found in response in APP/PS1 mice of different ages, compared with WT mice (Fig.4 D, F). The overall performance of the mice in the fear conditioning test revealed a signi cant decline in cognition capacity (working memory) of the APP/PS1 mice from the age of 4 months.

MiR-188-5p down-regulated the expression of Neuritin in 293T cells.
To validate the interaction of miR-188-5p with neuritin, we used the luciferase reporter constructs inserted with the 3'-UTR of neuritin and MUT-neuritin in 293T cells (Fig.5 A, B). As shown in Fig.5 C, the luciferase activity of neuritin 3'-UTR was higher than MUT-neuritin 3'-UTR, compared with the NC control and NC mimics (P<0.001).
To further identify the regulatory role of miR-188-5p in Neuritin expression, 293T cells were transfected with miR-188-5p mimics. Results from Western Blot analysis showed that, in protein level (Fig.5 D, E), the expression of Neuritin was signi cantly lower than NC mimics (P<0.05). Therefore, miR-188-5p could down-regulate the expression of Neuritin by binding to the neuritin 3'-UTR in 293T cells.
To evaluate the role of Neuritin in AD, we compared the expression level of Neuritin in the hippocampus of APP/PS1 mice at 1, 4, and 7 months with month-matched WT mice (Fig. 1). The assay at the mRNA level showed that, in all three month-groups, Neuritin was signi cantly reduced compared with WT mice in the hippocampus ( Fig. 1 A), especially at 4 months which was regarded as a sub-clinical or early-clinical stage of AD [34,35]. Similarly, the Western blot analysis showed that the expression level of the Neuritin protein was consistent with the observation at the Neuritin mRNA level (Fig. 1C, D).
MiRNAs are widely evidenced as key regulators of functions such as neurite outgrowth [36], dendritic spine morphology [37], and synaptic plasticity [38]. At present, functions and mechanisms of miRNAs in AD are increasingly recognized, which nding evidence about candidate miRNAs as biomarkers [39][40][41][42]. In order to explore the role of miRNAs and Neuritin in the pathogenesis of AD. We rst performed bioinformatics analysis in published datasets to cluster 2 candidate miRNAs that up-regulated in AD (Fig.   2C). Secondly, 4 miRNA target gene prediction websites with different algorithms were used to overlap 7 candidate miRNAs that may target the 3'-UTR of neuritin (Fig. 2D). Thirdly, multiple sequence homology analysis (Fig. 2E) and phylogenetic tree (Fig. 2F) of candidate miRNAs showed the identity of mmu-miR-188-5p and has-miR-188-5p, which corresponded to our previous studies [43], indicating a potential role of miR-188-5p in the pathogenesis of AD.
To further con rm the bioinformatics results, we detected the expression levels of miR-188-5p in the hippocampus and cerebral cortexes of the APP/PS1 mice. Consistent with the bioinformatics analysis, we found miR-188-5p is signi cantly high-expressed in the hippocampus, which plays one of the most essential roles in learning and memory [44], in 4 months APP/PS1 mice, compared with WT mice (Fig.  3A). Meanwhile, we detected the expression of Neuritin protein from the hippocampus of the same batch of mice (Fig. 3D). Interestingly, the expression level of miR-188-5p increased along with the decrease of the expression of Neuritin in the hippocampus at 4 months APP/PS1 mice (Fig. 3C, E). Revealing that change of Neuritin and miR-188-5p may be related to the genesis and development of AD.
Fear conditioning test is commonly used as a model for studying emotional learning and working memory in mice [45]. With the different brain circuitries of cognitive paradigms [46], we chose contextual and cued memory tests to evaluate the hippocampus-dependent [47] and amygdala dependent [48] memory in APP/PS1 mice (Fig. 4). The analysis of the freezing times from contextual and cued fear memory conditioning indicated that signi cant cognitive impairment occurred at the of 4 months and lasted until 7 months in APP/PS1 mice (Fig. 4C-F), notably in Short-term memory (STM) (Fig. 4B, P<0.001). It is well established that the earliest detectable de cits in cognition and neuropathological changes of AD are seen in the hippocampus, consistent with the progression of pathology [49,50]. Therefore, the results indicated that, with the appearance of cognitive impairment, the expression of Neuritin decreased, on the contrary, the expression of miR-188-5p increased, suggesting that miR-188-5p may be regulated the expression of Neuritin in the hippocampus of APP/PS1 mice.
Further, with the foundation data of the Fear conditioning test, we sought to determine the underlying mechanism of miR-188-5p, investigating whether miR-188-5p regulates the expression of Neuritin (Fig. 5). The luciferase reporter for Neuritin 3'-UTR was signi cantly inhibited by miR-188-5p (Fig. 5C), we got the further veri cation from the western blot of miR-188-5p mimics signi cantly inhibited the expression of Neuritin in 293T cells (Fig. 5D, E), further suggesting miR-188-5p may promote cognitive impairment through down-regulating the expression of Neuritin in the development of AD.

Conclusion
In summary, our study revealed a previously unde ned mechanism that miR-188-5p plays a signi cant role in the development of AD by inhibiting the expression of target protein Neuritin. These early alterations in miRNA and target protein in the hippocampus could be regarded as potential candidate biomarkers of AD. The effect of miR-188-5p and Neuritin on synaptic plasticity and pathology changes in AD needs to be explored in our future studies. Author contributions: JH, JZ and DS contributed to the conception or design of the study. YS and JS contributed to the acquisition, analysis, or interpretation of all cell-level data. GL and JS contributed to the acquisition, analysis, or interpretation of the AD mouse data. YS contributed to the acquisition, analysis and interpretation of the bioinformatics data. PM contributed to the construction of the vectors. JS and JH contributed to the drafting of the the article and revising it critically for all content. All authors contributed to revising the work critically for important intellectual content.

Abbreviations
Competing interests: The authors declare that they have no competing interests.
Availability of data and materials: The data used and/or analyzed during the present study are available from the corresponding author on reasonable request.
Con icts of interest: All experimental procedures were approved by the Animal Ethics Committee of Hangzhou Normal University.
Consent for publication: All participants provided written informed consent at enrolment including consent to publication.