Circulating Plasma microRNAs are Altered with Amyloidosis in a Mouse Model of Alzheimer’s Disease

Abstract

Pathological changes underlying Alzheimer’s disease (AD) begin decades before the classical symptoms of memory loss become evident. As microRNAs are released from neurons and enter the bloodstream, circulating microRNAs may be reflective of AD progression and are ideal candidates as biomarkers for early-stage disease detection. Here, we provide a novel, in-depth analysis of how plasma microRNAs alter with aging, the most prominent risk factor for AD, and with development of amyloid-β (Aβ) plaque deposition. We assessed the circulating microRNAs in APPswe/PSEN1dE9 transgenic mice and wild-type controls at 4, 8 and 15 m (n = 8–10) using custom designed Taqman arrays representing 185 neuropathology-related microRNAs. We performed a linear mixed-effects model to investigate the effects of age and genotype on plasma microRNAs expression. Following this analysis, we found 8 microRNAs were significantly affected by age alone in wild-type animals and 12 microRNAs altered in APPswe/PSEN1dE9 mice, either prior to Aβ plaque deposition (4 m) or during the development of AD-like pathogenesis (8 m or 15 m). Importantly, we found that differing sets of microRNAs were identified at each time point. Functional analysis of these data revealed that while common biological pathways, such as Inflammatory Response, were enriched throughout the disease process, Free Radical Scavenging, Immunological Disease, and Apoptosis Signaling were specifically enriched later in the disease process. Overall, this study reinforces that distinct biological processes underpin the early versus late stages of AD-like pathogenesis and highlights potential pre-symptomatic microRNAs biomarkers of neurodegeneration.

Keywords

Aging Alzheimer’s disease circulating microRNAs real-time polymerase chain reaction transgenic mice

INTRODUCTION

Alzheimer’s disease (AD), the most common form of dementia, is characterized by severe, progressive memory loss and impairment of cognitive function. The molecular basis of AD is clear in the rare, early-onset familial form of the disease, where mutations in either amyloid-β protein precursor (AβPP) or presenilin-1 or 2 (PSEN1, PSEN2), result in increased levels of the neurotoxic peptide, amyloid-β (Aβ) [1]. Sequential proteolytic cleavage of AβPP by β-site AβPP-cleaving enzyme 1 and PSEN/γ-secretase activity liberates Aβ. Although it is intensely debated whether increased levels of Aβ is the causative factor in the more common, sporadic late onset AD, both forms of the disease exhibit common clinical and pathological characteristics. A diagnosis of probable AD is currently achieved by psychometric tests such as Mini-Mental State Examination, which can be supported by structural or functional imaging [2 –4]. Conclusive diagnosis of the disease is only possible postmortem with the histological identification of the pathological features of Aβ plaques and hyperphosphorylated tau tangles [1].

Notably, AD has an extensive asymptomatic preclinical stage estimated to be as much as 15 to 20 years prior to clinical diagnosis [5]. During this phase, Aβ and hyperphosphorylated tau levels increase in the brain ultimately causing synaptic dysfunction, synaptic loss, and neurite retraction [6]. It has therefore been proposed that the most valuable therapeutic window occurs early in the disease process at a time when symptoms have yet to become evident. Consequently, discovery of an easily detectable biomarker of the molecular pathology underlying the early, preclinical stage of AD would be immensely valuable. Furthermore, increasing evidence points to differing processes distinguishing early and late stages of AD [7] and therefore, the identification of particular biomarkers reflective of pathology progression would also be of significant importance for future treatment strategies [8].

MicroRNA, small (18–25 nucleotides) non-coding RNA, have received considerable attention in recent AD research [9 –16]. They regulate protein synthesis in a tightly controlled and coordinated fashion by repressing translation or regulating degradation of target mRNA [17 –19]. An estimated half of all protein coding genes are regulated by microRNAs [20, 21], with both individual microRNA potentially targeting hundreds of different mRNA and multiple microRNAs capable of modulating a single mRNA [17]. MicroRNA-based mechanisms are important in learning, memory, and cognition [22] and there is growing evidence that dysregulation of microRNA may contribute to the onset and/or progression of AD by regulating disease-associated genes [15, 23]. MicroRNAs are released from neurons and glia in association with RNA-binding proteins, transport vesicles such as exosomes, microvesicles, and low-density lipoproteins [24]. They have shown a high level of stability over long time periods in blood fractions [25, 26] and as the blood absorbs cerebrospinal fluid, blood levels of microRNAs may closely reflect those of the brain [10 , 27–29]. This, alongside the evidence that brain microRNAs are dysregulated in AD, makes blood-borne microRNAs promising candidates as diagnostic tools for AD. Indeed a growing literature has identified sets of blood-borne microRNAs capable of differentiating cohorts of people with AD, mild cognitive impairment, and control cohorts [7 , 30–34], yet microRNA biomarkers associated with the preclinical phases and that can be used for monitoring the progression of the illness remain elusive. To begin to address this issue, we used a transgenic mouse model of AD (APPswe/PSEN1dE9, also described as APP/PS1). These mice recapitulate the early phenotype of the human disease by increasing the load of Aβ in the brain. They express a human/mouse chimeric AβPP and human presenilin-1, each carrying mutations associated with familial AD [35]. This leads to an early synaptic dysfunction [36] and chronic deposition of Aβ, neuroinflammation, and cognitive impairment from 6 m onwards [37 –39]. Furthermore, the majority of microRNAs appear to be evolutionarily conserved [40 –42] with high levels of expression conservation between mouse and human microRNA [43, 44]. Therefore, this transgenic mouse is a useful model to identify plasma microRNA changes occurring during early amyloidosis and the later stages of the illness. In this study, we aimed to understand the microRNA changes that occur through the aging process and in response to increased Aβ load, as these may provide critical insights into early pathogenesis and progression of AD.

MATERIALS AND METHODS

Animals

In this study, we used male litter-matched double transgenic APP/PS1 mice [37]. For details see the Jackson Laboratory (strain B6C3 Tg(APPswe, PSEN1dE9)85Dbo/J; http://jaxmice.jax.org/). These animals express chimeric mouse/ human APP695 with the Swedish mutation (K594M/N595 L) and the human exon-9-deleted variant of PSEN1 (PSEN1dE9) under independent mouse prion protein promoters.

All animal procedures for this study were carried out in accordance with the guidelines and approval of the University of Otago Animal Ethics Committee (approval number DET09/15). Animals were maintained as a colony at the University of Otago. Animals were group-housed in standard caging until 8 m of age when they were transferred to single housing to prevent injury from fighting between males. Food and water were available ad libitum, and the cage contained one red plastic tube (approximately 5 cm in diameter, 10 cm long) and shredded paper bedding as standard housing. Animals were kept on a 12 h light:dark cycle (lights on at 7 am), and the room temperature (RT) was controlled via a thermostat set at 21°C.

DNA extraction and genotyping

All mice aged 4, 8 and 15 m were genotyped for the presence of human exon-9-deleted variant PSEN1, which co-segregates with the APPswe gene. The DNA extraction protocol from mouse tail-tips was adapted from the Jackson Lab protocol, DNA from tail biopsies. In brief, mouse tail-tips were digested in DNA digestion buffer (100 nM Tris pH 8.5, 5 mM EDTA, 0.2 % SDS, 20 mM NaCl) with proteinase K (200μg/ml) and incubated overnight at 55°C with gentle shaking. Following brief agitation, samples were centrifuged at 12,000 g for 10 min at RT. The supernatant was isolated and an equal proportion of isopropanol added. Samples were mixed immediately to precipitate DNA and centrifuged (12,000 g, 10 min, RT). The resulting DNA pellet was resuspended in TE buffer (pH 8.0). Genotyping was performed by polymerase chain reaction (PCR) (according to the Jackson Lab protocol, Tg(PSEN1)) with primers specific to PSEN1 (Forward Sequence: AAT AGA GAA CGG CAG GAG CA, Reverse sequence: GCC ATG AGG GCA CTA ATC AT). Agarose gel electrophoresis stained with ethidium bromide showed either one band that indicated a wildtype animal or two bands indicating a transgenic animal.

Immunohistochemistry

Male APP/PS1 mice at 4, 8 and 15 m (n = 3 per group) were anesthetized with sodium pentobarbital and transcardially perfused with 4% chilled formaldehyde in 0.1 M phosphate buffer (pH 7.4). Brains were extracted and post-fixed in 4% paraformaldehyde (24 h, 4°C), then cryoprotected in 30% sucrose solution in 0.1 M phosphate buffer and subsequently serially sectioned (40μm). Coronal sections of the dorsal hippocampus (n = 3 biological replicates, with 5 technical replicates) were stained for Aβ plaques using Congo Red as described by Wilcock et al. [45] and counterstained with DAPI nuclear counterstain. In brief, sections were imaged with a fluorescence microscope (Zeiss) and analyzed using ImageJ [46]. Images were converted to 8 bit, a threshold value was determined and maintained for all images. An average hippocampal Aβ plaque burden was calculated as the percentage of the area of plaque deposition with respect to the total area assessed using the ImageJ algorithm.

Blood sampling, processing, and RNA isolation from plasma

Whole blood samples from male anesthetized transgenic mice and litter-matched wild-type controls (4, 8 and 15 m; n = 8–10) were collected by cardiac puncture during the light cycle. Immediately, samples were gently mixed in 0.5 ml EDTA vacutainer tubes (BD, New Jersey) and centrifuged (4°C, 15 min, 1200 g). The upper plasma layer was carefully isolated and then centrifuged (4°C, 15 min, 1200 g). All plasma samples were stored at –80°C. Total RNA was extracted using mirVana PARIS kit (Applied Biosystems) according to the manufacturer’s instructions. The concentration and purity were determined by spectrophotometry (A260, A260/A280 respectively; NanoDrop 1000 Spectrophotometer; NanoDrop Technologies, Waltham, MA).

MicroRNA profiling with custom TaqMan TLDAs

MicroRNAs were analyzed by quantitative reverse transcription PCR (qRT-PCR) using highly sensitive Taqman Low Density Array (TLDA) cards [47], custom designed to detect expression of neurodegeneration-associated microRNAs. A total of 185 relevant microRNAs and controls (see Supplementary Table 1) were chosen for profiling based on an in-depth analysis of microRNAs-related studies of age-related diseases including AD, Parkinson’s disease, mild cognitive impairment, and transgenic mouse models of AD.

Custom array cards were hybridized according to the manufacturer’s instructions with some modifications. Briefly, a fixed volume of extracted total RNA (3μl, 25 ng); A260/A280 ratio 1.9–2.1) was reverse transcribed to complementary DNA by using the TaqMan MicroRNAs RT Kit and microRNA-specific RT primers (Applied Biosystems). Pre-amplification reactions were performed with the TaqMan PreAmp Mastermix and custom Megaplex PreAmp microRNA Primer pools (Applied Biosystems) (Amplification cycles: 14). qRT-PCR reaction mixes were prepared with 4.5 μl of preamplification product and loaded into the array card ports (4 × 100 μl/ sample). To minimize technical variation, pairs of age and littermate matched wild-type control and transgenic samples were analyzed on a single TLDA. Array cards were run on an Applied Biosystems Vii7a Real-Time PCR system. Raw data were exported and Cycle passing Threshold (Ct) values were determined for each microRNA (baseline threshold set to “Automatic”) with Quantstudio Real-Time PCR software V1.2 (Applied Biosystems).

Quality Control, qPCR array filtering, and normalization

The quality of the samples were assessed using 1) expression of U6 snRNA controls to determine consistent RNA extraction; 2) expression of negative control ath-miR-159a to determine sample contamination; 3) ratio of miR-23a/miR-451 (ratios of <5 indicates low levels of hemolysis [48]). All plasma samples used for further analysis passed quality control assessments.

The Ct values from qPCR data were analyzed by using the open source HTqPCR v.1.26.0 Bioconductor software package [49] within the software environment R [50]. Briefly, all array data were imported into HTqPCR. The data were filtered to exclude microRNAs that were 1) assigned “Undetermined” by the qPCR software (Quantstudio); 2) exhibited Ct values <15 or >32; 3) control U6 snRNA; 4) negative control. The optimal normalization method was determined by assessing the dispersion of the data by measuring the coefficient of variation (CV) [51], following normalization by NormRank, Quantile, Scale Rank, Geometric mean [49] and deltadeltaCT (NormFinder [52]) (See Supplementary Table 2 and Supplementary Figure 1 for CV results and graphs). NormRank produced the lowest CV and therefore was used as the normalizing factor across all data, which also adjusted for minor differences in RNA recovery.

Statistical analysis

We fitted a linear mixed-effects model (lme4 (version 1.1-12), R (version 3.4.3) [53]) to investigate the relationship between microRNA expression and both age and genotype. This approach has been shown to appropriately model clustered data [54 –57]. In particular, some of the variance present in a data set can be attributed to random effects (e.g., litter effects). In our model, the dependent variable was normalized Ct value, with age and genotype as fixed effects (along with their interaction), and litter as a random effect. Assumptions of normality were checked by plotting a histogram of the standardized residuals and homogeneity of variance by plotting scatterplots of residuals against fitted values. We used Type II Wald chi square tests to determine if the fixed effects and their interaction in our model were significant. When significant main effects were detected, post-hoc Tukey’s Honest Significance Difference (HSD) tests were used to determine differences between groups, while adjusting for multiple testing, (p < 0.01).

Microarray data analysis

We utilized published, publicly available gene expression datasets from the Gene Expression Omnibus (GEO), a repository of high throughput gene expression data. We identified 3 gene expression datasets and compared brain gene expression between APP/PS1 mice and wild type controls at appropriate timepoints: 1) 4–5 months, Wt n = 11; APP/PS1 mice, n = 8, (accession number GSE74438-Illumina Arrays); 2) 8-9 months, Wt/APP/PS1 mice, n = 3, (accession number GSE104249-Affymetrix arrays); 3) 12 months, Wt/APP/PS1, n = 2, (accession number GSE87550-Illumina HiSeq). The data were analyzed with GEO2 R (an interactive data analysis tool: https://www.ncbi.nlm.nih.gov/geo), which uses limma (Linear Models for Microarray Analysis) R package to identify differentially expressed genes. Significance: p < 0.05).

MicroRNA target prediction, functional classification, and pathway analysis

MicroRNAs identified as differentially expressed at each time-point were further analyzed for putative target genes and pathways under their control. Ingenuity^® Pathway Analysis (IPA) (QIAGEN, http://www.ingenuity.com/products/ipa) was applied to identify biological functions and canonical pathways affected by the microRNAs altered with genotype. Putative target genes were identified utilizing the “microRNA Target Filter” tool in IPA. Only target genes which were predicted in silico with high confidence or were experimentally validated were included. This group was further filtered to include mRNA expressed in the mouse and human Nervous System (which included genes related to the “Disease” category: Immunological, Inflammatory, Cardiovascular, Neurological, Inflammatory response and Developmental disorder). Pathway enrichment analysis of these validated and putative target genes was performed by IPA’s “Core Analysis” tool. Network scores of 3 or above have at least a 99.9% likelihood of not being generated by chance. Gene-Enrichment/Functional Annotation Analysis was conducted using DAVID bioinformatics resources v6.8, as described previously [58, 59]. The significance of enrichment was determined using a modified Fisher’s exact test (Expression Analysis Systematic Explorer (EASE)) score ≤0.05).

Identification of validated microRNA-mRNA targets

Experimentally validated microRNAs-target interactions were identified by examining recent literature and querying the miRTarBase database (http://mirtarbase.mbc.nctu.edu.tw/php/index.php; release7.0). Only validation methods with strong evidence of interactions, including reporter assays, western blot or qPCR are referenced in this study.

RESULTS

Amyloid-β plaques are present from 8 m of age in APP/PS1 mice

We assessed the Aβ plaque load in hippocampal sections from transgenic and wild-type mice (4, 8 and 15 m; n = 3 per group). Aβ plaques, as identified by binding of Congo Red, were not detected in any wild-type samples (data not shown). By contrast, while no Aβ plaques were detected in the transgenic mice at 4 m, Aβ plaques were observed in the older groups (8 and 15 m; Fig. 1) confirming amyloidosis beginning between 4 and 8 m [37 , 61].

Fig.1

Age-related accumulation of Aβ plaques in the hippocampus of APP/PS1 transgenic mice. Representative examples of Congo red-stained hippocampal sections of APP/PS1 mice at 4 (A), 8 (B), and 15 (C) m. Blue counterstain = DAPI. D) Plaque load in the hippocampus, as determined by percentage of plaque area per total area measured, was abundant at 8 and 15 m (n = 3). Data are mean±s.e.m.

Plasma microRNA expression profiles are altered with aging

We first identified the effect of aging on microRNAs expression, specifically in the wild-type mice. Following the identification of microRNAs with significant effects for aging (Wald chi-squared tests p < 0.05) and post-hoc testing (Tukey’s HSD) to determine pairwise comparisons with time in the wild-type mice, we identified a total of 8 microRNAs altered with age (p < 0.01) (Table 1; Supplementary Figure 2). These microRNAs have been identified in previous aging-related studies of human or mouse circulating microRNAs [62 –69] (Supplementary Table 3). Interestingly, in our study, the levels of 5 microRNAs (miR-27b-3p, miR-143-3p, miR-361-5p, miR-382-5p, miR-423-5p) were found to be consistently lower at 15 m when compared to the values determined at 4 m (Table 1; Supplementary Figure 2: note that a lower normalized Ct = higher expression). By contrast, the levels of 3 microRNAs (miR-93-3p, miR-421-3p, miR-30d-5p) were significantly increased at 8 or 15 m. To examine the biological relevance of these plasma-derived microRNAs, we investigated whether any of these microRNAs had previously been associated with aging in mouse or human postmortem brain tissue. We found that all of the identified microRNAs were previously linked to aging in postmortem brain tissue (Supplementary Table 3: with associated references). Together, these data not only validate the alteration of plasma microRNAs with age, but also show that plasma microRNAs expression levels are dynamically affected by aging and may reflect changes within the brain.

Table 1

Temporal differential microRNAs expression in wild-type mice following Tukey post-hoc analysis

MicroRNA	4 m vs 8 m	4 m vs 8 m	4 m vs 15 m	4 m vs 15 m	8 m vs 15 m	8 m vs 15 m
	Fold Change	p-value	Fold Change	p-value	Fold Change	p-value
miR-27b-3p	−1.6	1.00E-03	−1.5	3.30E-03	1.1	NS
miR-30b-5p	0.7	NS	1.3	8.90E-03	1.2	NS
miR-93-3p	1.5	3.30E-03	1.2	NS	−1.2	NS
miR-143-3p	−1.5	NS	−1.8	1.60E-03	−1.2	NS
miR-361-5p	−1.1	NS	−2.0	1.30E-03	−1.8	NS
miR-382-5p	−1.6	NS	−2.2	8.50E-03	−1.3	NS
miR-421-3p	1.6	8.10E-03	1.3	NS	1.2	NS
miR-423-5p	-0.3	NS	−1.5	1.54E-03	1.4	NS

NS, not significant. Bold highlight, significant, p < 0.01, n = 10, 4 m and 15 m; n = 8, 8 m. Analysis was performed using Type II Wald chi square tests followed by post-hoc Tukey’s Honest Significance Difference (HSD) for each group, (4 m, 8 m and 15 m), (p < 0.01).

Plasma microRNA expression profiles are altered in APP/PS1 mice

We next investigated the effects of progressive AD-like pathology on plasma microRNA profiles. Following analysis of the linear mixed-effects model, we identified significant effects of age and genotype on the expression of 12 microRNAs including significant age x genotype interactions for 5 microRNAs (Table 2). Of the 12 identified microRNAs, 8 have been highlighted previously in AD-related studies of human or mouse circulating microRNAs [14 , 70–83] (Table 2).

Significant pairwise comparisons between these groups were identified using Tukey HSD post-hoc tests (p < 0.01, Fig. 2; Supplementary Figure 3). We found that at 4 m, a time representing the pre-pathological stages of AD, there was a significant decrease in expression of miR-200b-3p (–1.34 fold), miR-139-5p (–1.23 fold) and miR-27b-3p (–1.27 fold). In contrast, levels of miR-205-3p and miR-320-3p were significantly increased at 4 m (2 and 1.24 fold, respectively).

Table 2

Linear mixed model analysis results on the effects of age and genotype on microRNA expression

MicroRNA	Significance Model Effects			Previously Reported Changes with AD
	Age effect (p-value)	Genotype effect (p-value)	Age × Genotype Interaction (p-value)	Circulating AD	Post-mortem brain AD
miR-23b-3p	1.2E-04	7.8E-03	1.1E-01	Lugli et al. [72]	Cogswell et al. [71]
miR-27b-3p	3.8E-06	2.7E-05	8.3E-01	Guedes et al. [77]	Cogswell et al. [71]
				Cosin-Tomas et al. [70]
miR-34a-5p	1.1E-01	2.9E-02	5.3E-02	Kiko et al. [74]	Cogswell et al. [71]
				Satoh et al. [14]	Wang et al. [78, 79]
miR-139-5p	1.0E-02	5.7E-01	4.5E-02	Burgos et al. [29]	Berkis et al. [78, 79]
					Cogswell et al. [71]
miR-140-3p	6.6E-02	1.2E-05	5.6E-03	Dong et al. [75]	-
miR-143-3p	5.0E-04	8.0E-02	3.0E-02	Cheng et al. [76]	-
miR-200b-3p	6.0E-02	8.9E-01	4.0E-02	Guedes et al. [77]	Liu et al. [81]
miR-205-3p	7.2E-02	6.9E-03	2.2E-01	Cogswell et al. [71]	-
miR-320-3p	7.4E-01	2.1E-05	3.4E-01	Nagaraj et al. [34]	Herbert et al. [82]
miR-339-5p	1.7E-01	5.7E-02	1.4E-02	Ren et al. [73]	Cogswell et al. [71]
miR-486-3p	3.0E-03	7.6E-04	5.2E-01	-	-
miR-744-5p	3.5E-02	7.3E-04	5.4E-01	-	Lau et al. [83]

Significant Model Effects: Type II Wald chi-squared tests were used to determine if the fixed effects (age, genotype (wild-type and transgenic)) and their interaction in our linear mixed model were significant. (Bold highlight: p < 0.05). MicroRNAs changes reported previously in AD circulating studies and postmortem brain studies are shown. Circulating and postmortem brain AD studies found by literature search in Google Scholar with terms “MicroRNAs”, “miRNA”, “Alzheimer disease” AND “circulating”, “blood” OR “brain”, “transgenic mice”. Bold highlight: Transgenic mice studies.

Fig.2

A) Summary heatmap showing temporal expression profiles of significant microRNAs following Tukey post-hoc analyses, based on log2 mean Tg/Wt expression fold changes. Red-high expression; green-low expression. B) Expression fold changes of transgenic plasma microRNAs relative to wild-type controls. Bold highlight: Significant following post-hoc Tukey pair-wise analysis. n = 10 : 4 m and 15 m; n = 8 : 8 m.

At 8 m, when amyloidosis is apparent (Fig. 1), a different set of microRNAs were altered, with an increase in 4 microRNAs (miR-140-3p (1.63 fold), miR-486-3p (1.58 fold), miR-339-5p (1.31 fold) and miR-744-5p (1.36 fold)). We also found a decrease in 2 microRNAs, miR-143-3p (–1.39 fold), (which is apoptosis-related [84] and was also identified in our aging analysis) and miR-34a-5p (–1.58 fold). Our finding that miR-34a-5p was decreased, supports the recent proposal that this microRNA is a preclinical biomarker of AD [70].

At 15 m, miR-339-5p and miR-140-3p remained significantly increased (1.27 and 1.30 fold, respectively), suggesting a sustained increase in expression of these two microRNAs with time. Furthermore, miR-27b-3p, also identified in our aging analysis, was significantly decreased at 4 and 15 m (–1.28 at both times), with a modest decrease also observed at 8 m, supporting its role as an early stage marker of AD [34, 77]. In addition, miR-23b-3p, a direct modulator of autophagy-related gene expression [85] was significantly decreased (–1.91 fold).

To further explore the biological relevance of these plasma-derived microRNAs, we investigated whether expression changes in these microRNAs had been previously reported in studies of postmortem brain tissue from mouse models of AD or humans with AD. Nine of the 12 microRNAs were differentially expressed in AD or transgenic mouse postmortem tissue (Table 2). These results suggest that microRNAs signatures present in blood plasma are dynamically affected by AD-like pathogenesis over and above the effect of age alone, and that they may reflect pathological changes within the brain.

Plasma microRNAs altered in APP/PS1 mice may target pathways involved in AD

To explore the potential biological impact of the combined 12 microRNAs altered in the plasma of the APP/PS1 mice with amyloidosis, we identified brain-expressed mRNA targets of these microRNAs and used the pathway analysis tools of the IPA software to identify key relationships between the targets. Within the retrieved 1,190 predicted target genes, we found that genes associated with the physiological functions Organismal Survival and Cell Death and Survival formed the most significantly enriched Biological functions (Fig. 3a). Canonical Pathway analysis of this microRNA target gene set also revealed a number of enriched pathways (Fig. 4). Notably, these include G-Protein Coupled Receptor Signaling, Ephrin receptor Signaling, Axonal guidance Signaling, and numerous other pathways involved in learning and memory [86].

Fig.3

IPA Categorization of AD-related microRNAs target sets according to their Biological Functions using A) 1190 predicted mRNA targets of the 12 AD-related microRNAs and B) following analysis of the predicted target genes at each timepoint (260, 927, and 325 target genes, at 4, 8, and 15 m, respectively). The yellow line represents the threshold above which there are statistically significantly more genes in a biological function than expected by chance.

Fig.4

IPA Categorization of AD-related microRNAs sets according to their Canonical pathways using A) mRNA targets of the 12 AD-related microRNAs and B) following analysis of the predicted target genes from each age group. The yellow line represents the threshold above which there are statistically significantly more genes in a canonical pathway than expected by chance.

We next investigated whether the microRNA gene targets had similar or different biological functions at pre-versus post-plaque development. Following identification of the brain-expressed gene targets of the microRNAs altered at 4, 8 and 15 m (260, 927, and 325 target genes, respectively), we found significant overlap in the Biological Functions identified from our combined analysis (Fig. 3; Supplementary Figure 4). However, we also found that the function Free Radical Scavenging, a potential response to oxidative stress [87] was significantly enriched at 8 m and Immunological Disease was enriched at 8 and 15 m only, suggesting temporal effects on these functions that could be related to increased Aβ load (Fig. 3b). Similarly, we found that the majority of the identified Canonical Pathways overlapped with those seen following our combined analysis of the 12 microRNAs (Fig. 4; Supplementary Figure 5), however, Apoptosis Signaling which plays an important role in AD pathogenesis [88] was enriched at 8 and 15 m only (Fig. 4b), with the high level of enrichment seen at 8 m paralleling the onset of amyloidosis.

For a more comprehensive analysis of the microRNA gene targets at each timepoint, we investigated functionally related networks of genes and resulting regulatory hubs using IPA network analysis. At 4 m, the top network (Fig. 5A, Score 32) comprised several genes associated with the transforming growth factor-β (TGF-β) signaling pathway, which has a profound role in neuronal cell survival [89, 90]. These included SMAD genes, the main signal transducers for receptors of transforming growth factor-β (TGF-β) and TGFβR. Additionally, extracellular signal regulated kinases (ERK1/2) formed a hub within this network. ERK1/2 affect numerous physiological processes, but are extensively linked to synaptic plasticity and memory formation (reviewed in [91]). Consistent with these results, our analysis of GEO brain gene expression data from APP/PS1 transgenic mice compared to wild-type controls at 4 m, identified significant increases in the TGF-β signaling pathway genes found in this network (TGF-β, TGF-βR1, TGF-βR2, TGF-α), and significant decreases in Noggin and Smad6 (Fig. 5A).

Fig.5

IPA network analysis of AD-related microRNAs targets altered at (A) 4 m (Score 32); (microRNAs analyzed: miR-200b-3p; miR-139-5p; miR-320-3p; miR-205-3p; miR-27b-3p), (B) 8 m (Score 32); (microRNAs analyzed: miR-143-3p; miR-34a-5p; miR-486-3p; miR-744-5p; miR-339-5p; miR-140-3p), and (C) 15 m (Score 30); (microRNAs analyzed: miR-27b-3p; miR-339-5p; miR-140-3p; miR-23b-3p) following analysis of the predicted genes from each age group. Solid lines and dashed lines denote direct and indirect functional interaction of the products of the 2 genes, respectively. CP, canonical. D) Confirmation of changes in key IPA pathways and target genes in transgenic APP/PS1 mice versus wild-type controls following gene expression analysis using GEO. (i) TGF-β, (ii) TNF, and (iii) Apoptosis and ERK signaling related genes are significantly differentially expressed during pre-amyloid pathology (4–5 m), n = 8, early amyloid pathology (8–9 m), n = 3 and late amyloid pathology (12 m), n = 2. (Limma Analysis; p < 0.05); See text for further details.

At 8 m, our network analysis highlighted the major pro-inflammatory molecule Tumor Necrosis factor alpha (TNF) as a central hub (Fig. 5B, Score 32). Further supporting these results, we found that the expression of TNF associated genes, TNFRSF1A and TNFAIP2 and network genes FCGR2B, GUCY1A2, RASAL2, were altered in APP/PS1 mice compared to wild-type controls (Fig. 5B), following GEO brain gene expression data analysis at 8–9 m.

By 15 m, the top network (Fig. 5C, Score 30) comprised genes associated with Apoptosis signaling and ERK/MAPK signaling. The network genes (CASP2, CASP8, Casp9, SATB1, and Casp4) associated with these signaling pathways were also altered in APP/PS1 mice compared to wild-type controls (Fig. 5C) following GEO brain gene expression data analysis at 12 m.

In support of these results, a separate functional annotation clustering tool (DAVID analysis; Supplementary Table 4), also identified SMAD pathways (Enrichment score: 1.8) at 4 m (and 15 m) as well as Apoptosis related pathways at 8 and 15 m. Additionally, DAVID analysis highlighted enrichment of Cell Adhesion Molecules Cadherin Associated Proteins and Protein Phosphatases at 4 m, a number of protein kinases and Phospholipase at 8 m and Steroid Hormone Receptor Activity at 15 m (See Supplementary Table 4).

DISCUSSION

This study was designed to examine how plasma microRNAs alter with aging and increased Aβ plaque load in a mouse model of amyloidosis. We not only found changes in 8 aging-associated microRNAs, supporting previous studies, but identified changes in 12 plasma microRNAs both prior to and during amyloidosis in APP/PS1 mice. These data support the hypothesis that plasma microRNAs reflect molecular changes underpinning AD progression. Furthermore, differing sets of microRNAs were identified at each time point and the gene targets of these microRNAs were enriched in functional categories dysregulated in AD. The identification of plasma microRNAs altered early, pre-amyloid deposition, is a novel contribution from our study, as these microRNAs may be easily detectable surrogates for events such as inflammation in the AD brain [71 , 93–96]. Overall, this work provides critical insights into early pathogenesis and progression of AD.

MicroRNAs are master regulators of cellular processes and increasing evidence associates aging-related changes in gene expression with increased vulnerability to AD [97 –99]. Furthermore, their potential role as circulating biomarkers of the disease is gaining attention with a number of microRNAs now associated with the disease. Studies of human plasma have also provided evidence for age-associated changes in blood-borne microRNAs [100, 101]. Supporting the hypothesis that the plasma microRNAs are reflective of changes in the brain, we found that many microRNAs identified in APP/PS1 mouse plasma have previously been identified in postmortem studies of both AD and AD transgenic mouse models (Table 2). That the direction of change does not always precisely align is likely due to altered transport across the blood-brain barrier [102] and may be affected by inconsistent postmortem intervals [103].

The microRNAs identified in this study may contribute to a unique signature of amyloidosis, suitable to monitor AD progression and are strong targets for future study as preclinical biomarkers. In particular, those microRNA altered at 4 m (miR-200b-3p; miR-139-5p; miR-320-3p; miR-205-3p; miR-27b-3p) may be reflective of the very early disease process before pathological changes and cognitive symptoms are evident. Intriguingly, miR-200b was recently described as protective against Aβ peptide-induced toxicity [104], suggesting a significant role for this microRNA during the disease process, while miR-139-3p is known to modulate biological responses to proinflammatory stimuli [105]. MiR-27b-3p is involved in the regulation of autophagic clearance of damaged mitochondria [106], immune response [77] and regulates the anti-inflammatory transcription factor peroxisome proliferator-activated receptor γ (PPARγ) [107]. Importantly, miR-27b-3p and miR-200b-3p are both putative biomarkers of AD [81], with miR-27b-3p suggested as an early stage AD marker [34]. Here we find that miR-27b-3p is altered at a very early stage of the disease processes and so this microRNA may be of particular use as a pre-clinical biomarker of AD.

Interestingly, miR-27b-3p and miR-143-3p (altered at 8 m in APP/PS1 mice) were both highlighted in both our aging and amyloidosis analysis. MiR-143-3p has previously been shown to be involved in apoptosis and inflammatory TNF signaling, important functions identified in our IPA analyses and consistent with previous reports of mitochondrial impairment and inflammation altered with aging and AD [108 –111].

Functional analyses using the target genes of these AD-related microRNAs suggested early and sustained alterations of a number of common pathways including Inflammatory Response and others such as Cell Death and Survival, Molecular Transport, and Protein Synthesis (Fig. 3 and Supplementary Figure 4). Inflammation is believed to play a critical role in the development of AD. In particular, increasing levels of Aβ causes an induction of pro-inflammatory cytokines and microglia. The continuous Aβ aggregation does not allow the resolution of inflammation but leads to chronic inflammation which over time may cause distinct changes in the brain and lead ultimately to functional decline [112]. Our network (Fig. 5A) and GEO data analysis (which included 8 arrays per group), at the early time of 4 m, highlighted enrichment of TGF-β/SMAD signaling. Previous work has shown that TGF-βR1 and SMAD2 are targeted by miR-27b-3p, a regulatory microRNA of this network (validated by luciferase reporter assays and western blot) ([113], miRTarBase). TGF-β1 signaling is a key regulator of inflammatory processes [114, 115], including microglial activation [116] with insufficiency in TGF-β1/SMAD signaling an early event in AD pathogenesis.

Furthermore, increased neuroinflammation and TGF-β signaling in mild cognitively impaired and early AD patients occur before any evidence of Aβ plaques [114 , 117–120]. TGF-β has also been shown to protect against Aβ-induced neuroinflammation and apoptosis signaling [90]. It is possible that the unique set of microRNAs altered in APP/PS1 mouse plasma early at 4 m reflects an early neuroprotective response. There is mounting evidence that consequences of this chronic inflammation in AD may be alterations in brain metabolism occurring both before and after cognitive symptoms are evident in AD [121 –125]. Indeed, a recent study suggests that there may be an initial compensatory hypermetabolism (via inflammation, oxidative stress, and free radical formation) preceding hypometabolism occurring via diminished efficiency of energy production [126 –128]. The microRNAs altered at 8 and 15 m in the plasma may reflect a response to chronic amyloidosis via induced immune inflammatory response. This is supported by the interpretation of the most significant network formed at 8 m (Fig. 5B), which centers around TNF, a pro-inflammatory cytokine, with our GEO data analysis of brain gene expression also supporting changes in TNF related gene expression at this time. Furthermore, TNF is an experimentally validated target of miR-143-3p, a regulatory microRNA of this network (luciferase reporter assay and qPCR (miRTarBase)). Interestingly, this network also includes molecules involved in the regulation of apoptosis, a function also enriched in the network at 15 m (Fig. 5C).

A limitation of this study is that only male mice were assessed. As AD pathology is more prevalent in females (e.g., [129]) and there are sex differences in microRNA expression [130], a study including female mice is warranted. Our study intentionally focused on microRNAs known to be associated with neurodegeneration and present in plasma. The mice used in this study were litter-matched, utilizing longitudinal plasma samples from the same animals may have minimized experimental variation further. Future studies should examine microRNA expression patterns in human plasma and brain from early pre-symptomatic stages onwards.

One of the strengths of our study is the use of the TaqMan technology, which is both sensitive and specific [131, 132]. Our microRNAs fold changes range from 1.2 fold to over 3 fold. This level of expression change is expected from analysis of well-performed studies [133 –135], with larger fold changes suggesting poor normalization methodologies [132]. An earlier study of the temporal plasma microRNAs profiles of a triple transgenic mouse model of AD [136] do not overlap with those found in this study and this may be due to the differing AD mouse models, microRNAs analysis or statistical approaches used.

In conclusion, this study has confirmed alterations in recognized age-related microRNAs and thus, in doing so, confirmed the use of circulating microRNAs to monitor aging-related processes. It also shows that distinct age-related microRNAs patterns are evident in blood plasma during development of amyloidosis. This study not only highlight that plasma microRNAs levels are likely reflective of changes in the brain but show that their levels are dynamic, changing with the progression of AD. The unique sets of microRNAs identified here may prove useful as early pre-symptomatic biomarkers of AD and provide new insights into the development and progression of the disease.

Footnotes

ACKNOWLEDGMENTS

This study was supported by grants from Brain Research New Zealand, the Neurological Foundation of New Zealand and the Health Research Council of New Zealand. We thank Mr. Andrew Gray for his in-depth statistical advice and Dr. Shane Ohline for technical histochemistry assistance. We thank Mrs. Katie Peppercorn and Prof. Warren Tate for mouse genotyping.

Authors’ disclosures available online ().

The supplementary material is available in the electronic version of this article: .

References

Selkoe

, Hardy

(2016) The amyloid hypothesis of Alzheimer’s disease at 25 years. EMBO Mol Med 8, 595–608.

Zhou

, Wang

, Liu

, Yu

, Hu

, Chen

, Zeng

(2014) Advances of molecular imaging probes for the diagnosis of Alzheimer’s disease. Curr Alzheimer Res 11, 221–231.

Grundman

, Johnson

, Lu

, Siderowf

, Dell’Agnello

, Arora

, Skovronsky

, Mintun

, Pontecorvo

, 18F-AV-45-A17 Study Group (2016) Effect of amyloid imaging on the diagnosis and management of patients with cognitive decline: Impact of appropriate use criteria. Dement Geriatr Cogn Disord 41, 80–92.

Arevalo-Rodriguez

, Smailagic

, Roque

IFM

, Ciapponi

, Sanchez-Perez

, Giannakou

, Pedraza

, Bonfill Cosp

, Cullum

(2015) Mini-Mental State Examination (MMSE) for the detection of Alzheimer’s disease and other dementias in people with mild cognitive impairment (MCI). Cochrane Database Syst Rev CD010783.

Sperling

, Karlawish

, Johnson

(2013) Preclinical Alzheimer disease-the challenges ahead. Nat Rev Neurol 9, 54–58.

Mufson

, Ikonomovic

, Counts

, Perez

, Malek-Ahmadi

, Scheff

, Ginsberg

(2016) Molecular and cellular pathophysiology of preclinical Alzheimer’s disease. Behav Brain Res 311, 54–69.

Sheinerman

, Tsivinsky

, Abdullah

, Crawford

, Umansky

(2013) Plasma microRNA biomarkers for detection of mild cognitive impairment: Biomarker validation study. Aging 5, 925–938.

Sheinerman

, Tsivinsky

, Crawford

, Mullan

, Abdullah

, Umansky

(2012) Plasma microRNA biomarkers for detection of mild cognitive impairment. Aging 4, 590–605.

, Ong

, Seeher

, Armstrong

, Thalamuthu

, Brodaty

, Sachdev

, Mather

(2016) Circulating microRNAs as biomarkers of Alzheimer’s disease: A systematic review. J Alzheimers Dis 49, 755–766.

10.

Kumar

, Reddy

(2016) Are circulating microRNAs peripheral biomarkers for Alzheimer’s disease? Biochim Biophys Acta 1862, 1617–1627.

11.

Rao

, Benito

, Fischer

(2013) MicroRNAs as biomarkers for CNS disease. Front Mol Neurosci 6, 39.

12.

Batistela

, Josviak

, Sulzbach

, de Souza

(2017) An overview of circulating cell-free microRNAs as putative biomarkers in Alzheimer’s and Parkinson’s diseases. Int J Neurosci 127, 547–558.

13.

Leidinger

, Backes

, Deutscher

, Schmitt

, Mueller

, Frese

, Haas

, Ruprecht

, Paul

, Stahler

, Lang

, Meder

, Bartfai

, Meese

, Keller

(2013) A blood based 12-miRNA signature of Alzheimer disease patients. Genome Biol 14, R78.

14.

Satoh

, Kino

, Niida

(2015) MicroRNA-Seq data analysis pipeline to identify blood biomarkers for Alzheimer’s disease from public data. Biomark Insights 10, 21–31.

15.

Schonrock

, Matamales

, Ittner

, Gotz

(2012) MicroRNA networks surrounding APP and amyloid-beta metabolism-implications for Alzheimer’s disease. Exp Neurol 235, 447–454.

16.

Schonrock

, Ke

, Humphreys

, Staufenbiel

, Ittner

, Preiss

, Gotz

(2010) Neuronal microRNA deregulation in response to Alzheimer’s disease amyloid-beta. PLoS One 5, e11070.

17.

Maffioletti

, Tardito

, Gennarelli

, Bocchio-Chiavetto

(2014) Micro spies from the brain to the periphery: New clues from studies on microRNAs in neuropsychiatric disorders. Front Cell Neurosci 8, 75.

18.

, Coller

(2012) What comes first: Translational repression or mRNA degradation? The deepening mystery of microRNA function. Cell Res 22, 1322–1324.

19.

Gulyaeva

, Kushlinskiy

(2016) Regulatory mechanisms of microRNA expression. J Transl Med 14, 143.

20.

Krol

, Loedige

, Filipowicz

(2010) The widespread regulation of microRNA biogenesis, function and decay. Nat Rev Genet 11, 597–610.

21.

O’Carroll

, Schaefer

(2013) General principals of miRNA biogenesis and regulation in the brain. Neuropsychopharmacology 38, 39–54.

22.

Wang

, Kwon

, Tsai

(2012) MicroRNAs in learning, memory, and neurological diseases. Learn Mem 19, 359–368.

23.

Delay

, Mandemakers

, Hebert

(2012) MicroRNAs in Alzheimer’s disease. Neurobiol Dis 46, 285–290.

24.

Haqqani

, Delaney

, Tremblay

, Sodja

, Sandhu

, Stanimirovic

(2013) Method for isolation and molecular characterization of extracellular microvesicles released from brain endothelial cells. Fluids Barriers CNS 10, 4.

25.

Mitchell

, Parkin

, Kroh

, Fritz

, Wyman

, Pogosova-Agadjanyan

, Peterson

, Noteboom

, O’Briant

, Allen

, Lin

, Urban

, Drescher

, Knudsen

, Stirewalt

, Gentleman

, Vessella

, Nelson

, Martin

, Tewari

(2008) Circulating microRNAs as stable blood-based markers for cancer detection. Proc Natl Acad Sci U S A 105, 10513–10518.

26.

Turchinovich

, Weiz

, Langheinz

, Burwinkel

(2011) Characterization of extracellular circulating microRNA. Nucleic Acids Res 39, 7223–7233.

27.

Weiland

, Gao

, Zhou

, Mi

(2012) Small RNAs have a large impact: Circulating microRNAs as biomarkers for human diseases. RNA Biol 9, 850–859.

28.

Hoy

, Buck

(2012) Extracellular small RNAs: What, where, why? Biochem Soc Trans 40, 886–890.

29.

Burgos

, Malenica

, Metpally

, Courtright

, Rakela

, Beach

, Shill

, Adler

, Sabbagh

, Villa

, Tembe

, Craig

, Van Keuren-Jensen

(2014) Profiles of extracellular miRNA in cerebrospinal fluid and serum from patients with Alzheimer’s and Parkinson’s diseases correlate with disease status and features of pathology. PLoS One 9, e94839.

30.

Kumar

, Dezso

, MacKenzie

, Oestreicher

, Agoulnik

, Byrne

, Bernier

, Yanagimachi

, Aoshima

, Oda

(2013) Circulating miRNA biomarkers for Alzheimer’s disease. PLoS One 8, e69807.

31.

Sheinerman

, Toledo

, Tsivinsky

, Irwin

, Grossman

, Weintraub

, Hurtig

, Chen-Plotkin

, Wolk

, McCluskey

, Elman

, Trojanowski

, Umansky

(2017) Circulating brain-enriched microRNAs as novel biomarkers for detection and differentiation of neurodegenerative diseases. Alzheimers Res Ther 9, 89.

32.

Kumar

, Vijayan

, Bhatti

, Reddy

(2017) MicroRNAs as peripheral biomarkers in aging and age-related diseases. Prog Mol Biol Transl Sci 146, 47–94.

33.

Miya Shaik

, Tamargo

, Abubakar

, Kamal

, Greig

, Gan

(2018) The role of microRNAs in Alzheimer’s disease and their therapeutic potentials. Genes 9, E174.

34.

Nagaraj

, Laskowska-Kaszub

, Debski

, Wojsiat

, Dabrowski

, Gabryelewicz

, Kuznicki

, Wojda

(2017) Profile of 6 microRNA in blood plasma distinguish early stage Alzheimer’s disease patients from non-demented subjects. Oncotarget 8, 16122–16143.

35.

Jankowsky

, Slunt

, Ratovitski

, Jenkins

, Copeland

, Borchelt

(2001) Co-expression of multiple transgenes in mouse CNS: A comparison of strategies. Biomol Eng 17, 157–165.

36.

Ahmad

, Singh

, Das

, Gowaikar

, Shaw

, Ramachandran

, Rupanagudi

, Kommaddi

, Bennett

, Ravindranath

(2017) Reactive oxygen species-mediated loss of synaptic Akt1 signaling leads to deficient activity-dependent protein translation early in Alzheimer’s disease. Antioxid Redox Signal 27, 1269–1280.

37.

Jankowsky

, Fadale

, Anderson

, Xu

, Gonzales

, Jenkins

, Copeland

, Lee

, Younkin

, Wagner

, Younkin

, Borchelt

(2004) Mutant presenilins specifically elevate the levels of the 42 residue beta-amyloid peptide in vivo: Evidence for augmentation of a 42-specific gamma secretase. Hum Mol Genet 13, 159–170.

38.

Lee

, Han

(2013) An update of animal models of Alzheimer disease with a reevaluation of plaque depositions. Exp Neurobiol 22, 84–95.

39.

Savonenko

, Xu

, Melnikova

, Morton

, Gonzales

, Wong

, Price

, Tang

, Markowska

, Borchelt

(2005) Episodic-like memory deficits in the APPswe/PS1dE9 mouse model of Alzheimer’s disease: Relationships to beta-amyloid deposition and neurotransmitter abnormalities. Neurobiol Dis 18, 602–617.

40.

Lagos-Quintana

, Rauhut

, Meyer

, Borkhardt

, Tuschl

(2003) New microRNAs from mouse and human. RNA 9, 175–179.

41.

Lagos-Quintana

, Rauhut

, Lendeckel

, Tuschl

(2001) Identification of novel genes coding for small expressed RNAs. Science 294, 853–858.

42.

Bartel

, Chen

(2004) Micromanagers of gene expression: The potentially widespread influence of metazoan microRNAs. Nat Rev Genet 5, 396–400.

43.

Suryawanshi

, Vlad

, Lin

, Mantia-Smaldone

, Laskey

, Lee

, Lin

, Donnellan

, Klein-Patel

, Lee

, Mansuria

, Elishaev

, Budiu

, Edwards

, Huang

(2013) Plasma microRNAs as novel biomarkers for endometriosis and endometriosis-associated ovarian cancer. Clin Cancer Res 19, 1213–1224.

44.

Razak

, Ueno

, Takayama

, Nariai

, Nagasaki

, Saito

, Koso

, Lai

, Murakami

, Tsuji

, Michiue

, Nakauchi

, Otsu

, Watanabe

(2013) Profiling of microRNA in human and mouse ES and iPS cells reveals overlapping but distinct microRNA expression patterns. PLoS One 8, e73532.

45.

Wilcock

, Gordon

, Morgan

(2006) Quantification of cerebral amyloid angiopathy and parenchymal amyloid plaques with Congo red histochemical stain. Nat Protoc 1, 1591–1595.

46.

Schneider

, Rasband

, Eliceiri

(2012) NIH Image to ImageJ: 25 years of image analysis. Nat Methods 9, 671–675.

47.

Alvarez

, Done

(2014) SYBR(R) Green and Taq-Man(R) quantitative PCR arrays: Expression profile of genes relevant to a pathway or a disease state. Methods Mol Biol 1182, 321–359.

48.

Blondal

, Jensby Nielsen

, Baker

, Andreasen

, Mouritzen

, Wrang Teilum

, Dahlsveen

(2013) Assessing sample and miRNA profile quality in serum and plasma or other biofluids. Methods 59, S1–S6.

49.

Dvinge

, Bertone

(2009) HTqPCR: High-throughput analysis and visualization of quantitative real-time PCR data in R. Bioinformatics 25, 3325–3326.

50.

Gentleman

, Carey

, Bates

, Bolstad

, Dealing

, Dudoit

, Ellis

, Gautier

, Ge

, Gentry

, Hornik

, Hothorn

, Huber

, Iacus

, Irizarry

, Leisch

, Li

, Maechler

, Rossini

, Sawitzki

, Smith

, Smyth

, Tierney

, Yang

, Zhang

(2004) Bioconductor: Open software development for computational biology and bioinformatics. Genome Biol 5. R80.

51.

Deo

, Carlsson

, Lindlof

(2011) How to choose a normalization strategy for miRNA quantitative real-time (qPCR) arrays. J Bioinform Comput Biol 9, 795–812.

52.

Andersen

, Jensen

, Orntoft

(2004) Normalization of real-time quantitative reverse transcription-PCR data: A model-based variance estimation approach to identify genes suited for normalization, applied to bladder and colon cancer data sets. Cancer Res 64, 5245–5250.

53.

Bates

, Mächler

, Bolker

, Walker

, (2015) lme4: Linear mixed-effects models using Eigen and S4. R package version 1.1-8, http://CRAN.R-project.org/package=lme4.

54.

Huan

, Chen

, Liu

, Bhattacharya

, Rong

, Chen

, Seshadri

, Tanriverdi

, Freedman

, Larson

, Murabito

, Levy

(2018) Age-associated microRNA expression in human peripheral blood is associated with all-cause mortality and age-related traits. Aging Cell 17, e12687.

55.

Meliza

, Margoliash

(2012) Emergence of selectivity and tolerance in the avian auditory cortex. J Neurosci 32, 15158–15168.

56.

Steibel

, Poletto

, Coussens

, Rosa

(2009) A powerful and flexible linear mixed model framework for the analysis of relative quantification RT-PCR data. Genomics 94, 146–152.

57.

Moen

, Fricano-Kugler

, Luikart

, O’Malley

(2016) Analyzing clustered data: Why and how to account for multiple observations nested within a study participant? PLoS One 11, e0146721.

58.

da Huang

, Sherman

, Lempicki

(2009) Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources. Nat Protoc 4, 44–57.

59.

da Huang

, Sherman

, Lempicki

(2009) Bioinformatics enrichment tools: Paths toward the comprehensive functional analysis of large gene lists. Nucleic Acids Res 37, 1–13.

60.

Gordon

, Holcomb

, Jantzen

, DiCarlo

, Wilcock

, Boyett

, Connor

, Melachrino

, O’Callaghan

, Morgan

(2002) Time course of the development of Alzheimer-like pathology in the doubly transgenic PS1+APP mouse. Exp Neurol 173, 183–195.

61.

Garcia-Alloza

, Robbins

, Zhang-Nunes

, Purcell

, Betensky

, Raju

, Prada

, Greenberg

, Bacskai

, Frosch

(2006) Characterization of amyloid deposition in the APPswe/PS1dE9 mouse model of Alzheimer disease. Neurobiol Dis 24, 516–524.

62.

Dhahbi

, Spindler

, Atamna

, Yamakawa

, Guerrero

, Boffelli

, Mote

, Martin

(2013) Deep sequencing identifies circulating mouse miRNAs that are functionally implicated in manifestations of aging and responsive to calorie restriction. Aging 5, 130–141.

63.

ElSharawy

, Keller

, Flachsbart

, Wendschlag

, Jacobs

, Kefer

, Brefort

, Leidinger

, Backes

, Meese

, Schreiber

, Rosenstiel

, Franke

, Nebel

(2012) Genome-wide miRNA signatures of human longevity. Aging Cell 11, 607–616.

64.

Freedman

, Gerstein

, Mick

, Rozowsky

, Levy

, Kitchen

, Das

, Shah

, Danielson

, Beaulieu

, Navarro

, Wang

, Galeev

, Holman

, Kwong

, Murthy

, Tanriverdi

, Koupenova-Zamor

, Mikhalev

, Tanriverdi

(2016) Diverse human extracellular RNAs are widely detected in human plasma. Nat Commun 7, 11106.

65.

Inukai

, de Lencastre

, Turner

, Slack

(2012) Novel microRNAs differentially expressed during aging in the mouse brain. PLoS One 7, e40028.

66.

, Bates

, An

, Terry

, Wang

(2011) Up-regulation of key microRNAs, and inverse down-regulation of their predicted oxidative phosphorylation target genes, during aging in mouse brain. Neurobiol Aging 32, 944–955.

67.

Mohammed

, Rhee

, Phee

, Kim

, Lee

, Park

, Jung

, Kim

, Park

, Nam

, Kim

(2016) miR-204 downregulates EphB2 in aging mouse hippocampal neurons. Aging Cell 15, 380–388.

68.

Olivieri

, Spazzafumo

, Santini

, Lazzarini

, Albertini

, Rippo

, Galeazzi

, Abbatecola

, Marcheselli

, Monti

, Ostan

, Cevenini

, Antonicelli

, Franceschi

, Procopio

(2012) Age-related differences in the expression of circulating microRNAs: MiR-21 as a new circulating marker of inflammaging. Mech Ageing Dev 133, 675–685.

69.

Wang

, Huang

, Hu

, Stromberg

, Nelson

(2011) Patterns of microRNA expression in normal and early Alzheimer’s disease human temporal cortex: White matter versus gray matter. Acta Neuropathol 121, 193–205.

70.

Cosin-Tomas

, Antonell

, Llado

, Alcolea

, Fortea

, Ezquerra

, Lleo

, Marti

, Pallas

, Sanchez-Valle

, Molinuevo

, Sanfeliu

, Kaliman

(2017) Plasma miR-34a-5p and miR-545-3p as early biomarkers of Alzheimer’s disease: Potential and limitations. Mol Neurobiol 54, 5550–5562.

71.

Cogswell

, Ward

, Taylor

, Waters

, Shi

, Cannon

, Kelnar

, Kemppainen

, Brown

, Chen

, Prinjha

, Richardson

, Saunders

, Roses

, Richards

(2008) Identification of miRNA changes in Alzheimer’s disease brain and CSF yields putative biomarkers and insights into disease pathways. J Alzheimers Dis 14, 27–41.

72.

Lugli

, Cohen

, Bennett

, Shah

, Fields

, Hernandez

, Smalheiser

(2015) Plasma exosomal miRNAs in persons with and without Alzheimer disease: Altered expression and prospects for biomarkers. PLoS One 10, e0139233.

73.

Ren

, Zhang

, Dammer

, Zhou

, Wang

, Liu

, Feng

, Jiang

, Chen

, Wang

, Cheng

(2016) Peripheral blood microrna expression profiles in Alzheimer’s disease: Screening, validation, association with clinical phenotype and implications for molecular mechanism. Mol Neurobiol 53, 5772–5781.

74.

Kiko

, Nakagawa

, Tsuduki

, Furukawa

, Arai

, Miyazawa

(2014) MicroRNAs in plasma and cerebrospinal fluid as potential markers for Alzheimer’s disease. J Alzheimers Dis 39, 253–259.

75.

Dong

, Li

, Huang

, Chen

, Li

, Wang

, Hu

, Xu

, Zhang

, Zen

, Xiao

, Yan

, Wang

, Zhang

(2015) Serum microRNA profiles serve as novel biomarkers for the diagnosis of Alzheimer’s disease. Dis Markers 2015, 625659.

76.

Cheng

, Doecke

, Sharples

, Villemagne

, Fowler

, Rembach

, Martins

, Rowe

, Macaulay

, Masters

, Hill

, Australian Imaging Biomarkers and Lifestyle (AIBL) Research Group (2015) Prognostic serum miRNA biomarkers associated with Alzheimer’s disease shows concordance with neuropsychological and neuroimaging assessment. Mol Psychiatry 20, 1188–1196.

77.

Guedes

, Santana

, Cunha

, Duro

, Almeida

, Cardoso

, de Lima

, Cardoso

(2016) MicroRNA deregulation and chemotaxis and phagocytosis impairment in Alzheimer’s disease. Alzheimers Dement (Amst) 3, 7–17.

78.

Wang

, Liu

, Zhu

, Xu

, Ma

, Dai

, Huang

, Liu

, Zhang

, Qin

(2009) miR-34a, a microRNA up-regulated in a double transgenic mouse model of Alzheimer’s disease, inhibits bcl2 translation. Brain Res Bull 80, 268–273.

79.

Wang

, Min

, Guo

, Zhang

, Jiang

, Shao

, Xing

, Yin

, Liu

, Guo

(2017) Profiling microRNA from brain by microarray in a transgenic mouse model of Alzheimer’s disease. Biomed Res Int 2017, 8030369.

80.

Bekris

, Lutz

, Montine

, Yu

, Tsuang

, Peskind

, Leverenz

(2013) MicroRNA in Alzheimer’s disease: An exploratory study in brain, cerebrospinal fluid and plasma. Biomarkers 18, 455–466.

81.

Liu

, Wang

, Li

, Xue

, Zhang

, Wang

(2014) MicroRNA-135a and -200b, potential Biomarkers for Alzheimers disease, regulate beta secretase and amyloid precursor protein. Brain Res 1583, 55–64.

82.

Hebert

, Horre

, Nicolai

, Papadopoulou

, Mandemakers

, Silahtaroglu

, Kauppinen

, Delacourte

, De Strooper

(2008) Loss of microRNA cluster miR-29a/b-1 in sporadic Alzheimer’s disease correlates with increased BACE1/beta-secretase expression. Proc Natl Acad Sci U S A 105, 6415–6420.

83.

Lau

, Bossers

, Janky

, Salta

, Frigerio

, Barbash

, Rothman

, Sierksma

, Thathiah

, Greenberg

, Papadopoulou

, Achsel

, Ayoubi

, Soreq

, Verhaagen

, Swaab

, Aerts

, De Strooper

(2013) Alteration of the microRNA network during the progression of Alzheimer’s disease. EMBO Mol Med 5, 1613–1634.

84.

Yang

, Wang

, Pan

, Zhang

(2018) miR-143-3p regulates cell proliferation and apoptosis by targeting IGF1R and IGFBP5 and regulating the Ras/p38 MAPK signaling pathway in rheumatoid arthritis. Exp Ther Med 15, 3781–3790.

85.

Sun

, Liu

, Zhang

, Ji

, Li

, Yang

, Wu

, Guo

(2016) miR-23b improves cognitive impairments in traumatic brain injury by targeting ATG12-mediated neuronal autophagy. Behav Brain Res 340, 126–136.

86.

Gillingwater

, Wishart

(2013) Mechanisms underlying synaptic vulnerability and degeneration in neurodegenerative disease. Neuropathol Appl Neurobiol 39, 320–334.

87.

Ahmadinejad

, Geir Moller

, Hashemzadeh-Chaleshtori

, Bidkhori

, Jami

(2017) Molecular mechanisms behind free radical scavengers function against oxidative stress. Antioxidants (Basel) 6, E51.

88.

Obulesu

, Lakshmi

(2014) Apoptosis in Alzheimer’s disease: An understanding of the physiology, pathology and therapeutic avenues. Neurochem Res 39, 2301–2312.

89.

Tesseur

, Zou

, Esposito

, Bard

, Berber

, Can

, Lin

, Crews

, Tremblay

, Mathews

, Mucke

, Masliah

, Wyss-Coray

(2006) Deficiency in neuronal TGF-beta signaling promotes neurodegeneration and Alzheimer’s pathology. J Clin Invest 116, 3060–3069.

90.

Diniz

, Tortelli

, Matias

, Morgado

, Bergamo Araujo

, Melo

, Seixas da Silva

, Alves-Leon

, de Souza

, Ferreira

, De Felice

, Gomes

FCA

(2017) Astrocyte transforming growth factor beta 1 protects synapses against Abeta oligomers in Alzheimer’s disease model. J Neurosci 37, 6797–6809.

91.

Teich

, Nicholls

, Puzzo

, Fiorito

, Purgatorio

, Fa

, Arancio

(2015) Synaptic therapy in Alzheimer’s disease: A CREB-centric approach. Neurotherapeutics 12, 29–41.

92.

Michaud

, Balardy

, Moulis

, Gaudin

, Peyrot

, Vellas

, Cesari

, Nourhashemi

(2013) Proinflammatory cytokines, aging, and age-related diseases. J Am Med Dir Assoc 14, 877–882.

93.

Liu

, Tian

, Ander

, Xu

, Stamova

, Zhan

, Turner

, Jickling

, Sharp

(2010) Brain and blood microRNA expression profiling of ischemic stroke, intracerebral hemorrhage, and kainate seizures. J Cereb Blood Flow Metab 30, 92–101.

94.

Chatterjee

, Roy

(2017) Comparative analysis of RNA-Seq data from brain and blood samples of Parkinson’s disease. Biochem Biophys Res Commun 484, 557–564.

95.

Jeyaseelan

, Lim

, Armugam

(2008) MicroRNA expression in the blood and brain of rats subjected to transient focal ischemia by middle cerebral artery occlusion. Stroke 39, 959–966.

96.

Geekiyanage

, Jicha

, Nelson

, Chan

(2012) Blood serum miRNA: Non-invasive biomarkers for Alzheimer’s disease. Exp Neurol 235, 491–496.

97.

Avramopoulos

, Szymanski

, Wang

, Bassett

(2011) Gene expression reveals overlap between normal aging and Alzheimer’s disease genes. Neurobiol Aging 32, 2319 e2327–e2334.

98.

Fernandes

, Wan

, Tacutu

, Barardo

, Rajput

, Wang

, Thoppil

, Thornton

, Yang

, Freitas

, de Magalhaes

(2016) Systematic analysis of the gerontome reveals links between aging and age-related diseases. Hum Mol Genet 25, 4804–4818.

99.

Reddy

, Tonk

, Kumar

, Vijayan

, Kandimalla

, Kuruva

, Reddy

(2016) A critical evaluation of neuroprotective and neurodegenerative MicroRNAs in Alzheimer’s disease. Biochem Biophys Res Commun 483, 1156–1165.

100.

Noren Hooten

, Abdelmohsen

, Gorospe

, Ejiogu

, Zonderman

, Evans

(2010) microRNA expression patterns reveal differential expression of target genes with age. PLoS One 5, e10724.

101.

Ameling

, Kacprowski

, Chilukoti

, Malsch

, Liebscher

, Suhre

, Pietzner

, Friedrich

, Homuth

, Hammer

, Volker

(2015) Associations of circulating plasma microRNAs with age, body mass index and sex in a population-based study. BMC Med Genomics 8, 61.

102.

Cheng

, Quek

, Sun

, Bellingham

, Hill

(2013) The detection of microRNA associated with Alzheimer’s disease in biological fluids using next-generation sequencing technologies. Front Genet 4, 150.

103.

Sethi

, Lukiw

(2009) Micro-RNA abundance and stability in human brain: Specific alterations in Alzheimer’s disease temporal lobe neocortex. Neurosci Lett 459, 100–104.

104.

Higaki

, Muramatsu

, Matsuda

, Matsumoto

, Satoh

, Michikawa

, Niida

(2018) Defensive effect of microRNA-200b/c against amyloid-beta peptide-induced toxicity in Alzheimer’s disease models. PLoS One 13, e0196929.

105.

Tang

, Bao

, Su

, Huang

(2017) MicroRNA-139 modulates Alzheimer’s-associated pathogenesis in SAMP8 mice by targeting cannabinoid receptor type 2. Genet Mol Res 16.

106.

Kim

, Fiesel

, Belmonte

, Hudec

, Wang

, Kim

, Nelson

, Springer

, Kim

(2016) miR-27a and miR-27b regulate autophagic clearance of damaged mitochondria by targeting PTEN-induced putative kinase 1 (PINK1). Mol Neurodegener 11, 55.

107.

Jennewein

, von Knethen

, Schmid

, Brune

(2010) MicroRNA-27b contributes to lipopolysaccharide-mediated peroxisome proliferator-activated receptor gamma (PPARgamma) mRNA destabilization. J Biol Chem 285, 11846–11853.

108.

Grimm

, Eckert

(2017) Brain aging and neurodegeneration: From a mitochondrial point of view. J Neurochem 143, 418–431.

109.

Vallee

, Lecarpentier

(2016) Alzheimer disease: Crosstalk between the canonical Wnt/beta-catenin pathway and PPARs alpha and gamma. Front Neurosci 10, 459.

110.

Chung

, Seo

, Chung

, Kim

, Leeuwenburgh

, Yu

, Chung

(2008) Molecular mechanism of PPAR in the regulation of age-related inflammation. Ageing Res Rev 7, 126–136.

111.

Liu

, Du

, Zhu

, Zhang

, Sun

, Liu

, Li

, Wang

, Du

(2015) Identification of apoptosis-related microRNAs and their target genes in myocardial infarction post-transplantation with skeletal myoblasts. J Transl Med 13, 270.

112.

Heneka

, Golenbock

, Latz

(2015) Innate immunity in Alzheimer’s disease. Nat Immunol 16, 229–236.

113.

Zeng

, Huang

, Senavirathna

, Wang

, Liu

(2017) miR-27b inhibits fibroblast activation via targeting TGFbeta signaling pathway. BMC Cell Biol 18, 9.

114.

Caraci

, Battaglia

, Bruno

, Bosco

, Carbonaro

, Giuffrida

, Drago

, Sortino

, Nicoletti

, Copani

(2011) TGF-beta1 pathway as a new target for neuroprotection in Alzheimer’s disease. CNS Neurosci Ther 17, 237–249.

115.

Chen

, Ke

, Lu

, Qiu

, Peng

(2015) Protection of TGF-beta1 against neuroinflammation and neurodegeneration in Abeta1-42-induced Alzheimer’s disease model rats. PLoS One 10, e0116549.

116.

Brionne

, Tesseur

, Masliah

, Wyss-Coray

(2003) Loss of TGF-beta 1 leads to increased neuronal cell death and microgliosis in mouse brain. Neuron 40, 1133–1145.

117.

Tarkowski

, Andreasen

, Tarkowski

, Blennow

(2003) Intrathecal inflammation precedes development of Alzheimer’s disease. J Neurol Neurosurg Psychiatry 74, 1200–1205.

118.

Huang

, Jia

, Liu

(2013) Decreased serum levels of the angiogenic factors VEGF and TGF-beta1 in Alzheimer’s disease and amnestic mild cognitive impairment. Neurosci Lett 550, 60–63.

119.

von Bernhardi

, Cornejo

, Parada

, Eugenin

(2015) Role of TGFbeta signaling in the pathogenesis of Alzheimer’s disease. Front Cell Neurosci 9, 426.

120.

Haque

, Nazir

(2016) SMAD Transcription factor, Sma-9, attunes TGF-beta signaling cascade towards modulating amyloid beta aggregation and associated outcome in transgenic C. elegans. Mol Neurobiol 53, 109–119.

121.

Insel

, Ossenkoppele

, Gessert

, Jagust

, Landau

, Hansson

, Weiner

, Mattsson

Alzheimer’s Disease Neuroimaging Initiative (2017) Time to amyloid positivity and preclinical changes in brain metabolism, atrophy, and cognition: Evidence for emerging amyloid pathology in Alzheimer’s disease. Front Neurosci 11, 281.

122.

Landau

, Mintun

, Joshi

, Koeppe

, Petersen

, Aisen

, Weiner

, Jagust

, Alzheimer’s Disease Neuroimaging Initiative (2012) Amyloid deposition, hypometabolism, and longitudinal cognitive decline. Ann Neurol 72, 578–586.

123.

Wirth

, Bejanin

, La Joie

, Arenaza-Urquijo

, Gonneaud

, Landeau

, Perrotin

, Mezenge

, de La Sayette

, Desgranges

, Chetelat

(2018) Regional patterns of gray matter volume, hypometabolism, and beta-amyloid in groups at risk of Alzheimer’s disease. Neurobiol Aging 63, 140–151.

124.

Takkinen

, Lopez-Picon

, Al Majidi

, Eskola

, Krzyczmonik

, Keller

, Loyttyniemi

, Solin

, Rinne

, Haaparanta-Solin

(2017) Brain energy metabolism and neuroinflammation in ageing APP/PS1-21 mice using longitudinal (18)F-FDG and (18)F-DPA-714 PET imaging. J Cereb Blood Flow Metab 37, 2870–2882.

125.

Yin

, Sancheti

, Patil

, Cadenas

(2016) Energy metabolism and inflammation in brain aging and Alzheimer’s disease. Free Radic Biol Med 100, 108–122.

126.

Ashraf

, Fan

, Brooks

, Edison

(2015) Cortical hypermetabolism in MCI subjects: A compensatory mechanism? Eur J Nucl Med Mol Imaging 42, 447–458.

127.

Do Carmo

, Crynen

, Paradis

, Reed

, Iulita

, Ducatenzeiler

, Crawford

, Cuello

(2017) Hippocampal proteomic analysis reveals distinct pathway deregulation profiles at early and late stages in a rat model of Alzheimer’s-like amyloid pathology. Mol Neurobiol 55, 3451–3476.

128.

Takano

, Yamashita

, Nagano

, Otani

, Maekura

, Kamada

, Tsunoda

, Tsutsumi

, Tomiyama

, Mori

, Matsuura

, Matsuyama

(2013) Proteomic analysis of the hippocampus in Alzheimer’s disease model mice by using two-dimensional fluorescence difference in gel electrophoresis. Neurosci Lett 534, 85–89.

129.

Alberca

, Montes-Latorre

, Gil-Neciga

, Mir-Rivera

, Lozano-San Martin

(2002) Alzheimer’s disease and women. Rev Neurol 35, 571–579.

130.

Guo

, Zhang

, Ma

, Wang

, Liang

(2017) miRNA and mRNA expression analysis reveals potential sex-biased miRNA expression. Sci Rep 7, 39812.

131.

Mestdagh

, Hartmann

, Baeriswyl

, Andreasen

, Bernard

, Chen

, Cheo

, D’Andrade

, DeMayo

, Dennis

, Derveaux

, Feng

, Fulmer-Smentek

, Gerstmayer

, Gouffon

, Grimley

, Lader

, Lee

, Luo

, Mouritzen

, Narayanan

, Patel

, Peiffer

, Ruberg

, Schroth

, Schuster

, Shaffer

, Shelton

, Silveria

, Ulmanella

, Veeramachaneni

, Staedtler

, Peters

, Guettouche

, Wong

, Vandesompele

(2014) Evaluation of quantitative miRNA expression platforms in the microRNA quality control (miRQC) study. Nat Methods 11, 809–815.

132.

Witwer

, Halushka

(2016) Toward the promise of microRNAs - Enhancing reproducibility and rigor in microRNA research. RNA Biol 13, 1103–1116.

133.

Galimberti

, Villa

, Fenoglio

, Serpente

, Ghezzi

, Cioffi

, Arighi

, Fumagalli

, Scarpini

(2014) Circulating miRNAs as potential biomarkers in Alzheimer’s disease. J Alzheimers Dis 42, 1261–1267.

134.

Krzeszinski

, Wei

, Huynh

, Jin

, Wang

, Chang

, Xie

, He

, Mangala

, Lopez-Berestein

, Sood

, Mendell

, Wan

(2014) miR-34a blocks osteoporosis and bone metastasis by inhibiting osteoclastogenesis and Tgif2. Nature 512, 431–435.

135.

Sorensen

, Nygaard

, Christensen

(2016) miRNA expression profiles in cerebrospinal fluid and blood of patients with Alzheimer’s disease and other types of dementia - an exploratory study. Transl Neurodegener 5, 6

136.

Garza-Manero

, Arias

, Bermudez-Rattoni

, Vaca

, Zepeda

(2015) Identification of age- and disease-related alterations in circulating miRNAs in a mouse model of Alzheimer’s disease. Front Cell Neurosci 9, 53.