MiRNA-15b and miRNA-125b are associated with regional Aβ-PET and FDG-PET uptake in cognitively normal individuals with subjective memory complaints

There is substantial experimental evidence for dysregulation of several microRNA (miRNA) expression levels in Alzheimer’s disease (AD). MiRNAs modulate critical brain intracellular signaling pathways and are associated with AD core pathophysiological mechanisms. First, we conducted a real-time quantitative PCR-based pilot study to identify a set of brain-enriched miRNAs in a monocentric cohort of cognitively normal individuals with subjective memory complaints, a condition associated with increased risk of AD. Second, we investigated the impact of age, sex, and the Apolipoprotein E ε4 (APOE ε4) allele, on the identified miRNA plasma concentrations. In addition, we explored the cross-sectional and longitudinal association of the miRNAs plasma concentrations with regional brain metabolic uptake using amyloid-β (Aβ)-positron emission tomography (Aβ-PET) and 18F-fluorodeoxyglucose-PET (18F-FDG-PET). We identified a set of six brain-enriched miRNAs—miRNA-125b, miRNA-146a, miRNA-15b, miRNA-148a, miRNA-26b, and miRNA-100. Age, sex, and APOE ε4 allele were not associated with individual miRNA abundance. MiRNA-15b concentrations were significantly lower in the Aβ-PET-positive compared to Aβ-PET-negative individuals. Furthermore, we found a positive effect of the miRNA-15b*time interaction on regional metabolic 18F-FDG-PET uptake in the left hippocampus. Plasma miRNA-125b concentrations, as well as the miRNA-125b*time interaction (over a 2-year follow-up), were negatively associated with regional Aβ-PET standard uptake value ratio in the right anterior cingulate cortex. At baseline, we found a significantly negative association between plasma miRNA-125b concentrations and 18F-FDG-PET uptake in specific brain regions. In an asymptomatic at-risk population for AD, we show significant associations between plasma concentrations of miRNA-125b and miRNA-15b with core neuroimaging biomarkers of AD pathophysiology. Our results, coupled with existing experimental evidence, suggest a potential protective anti-Aβ effect of miRNA-15b and a biological link between miRNA-125b and Aβ-independent neurotoxic pathways.


Introduction
MicroRNAs (miRNAs)-a class of endogenous, singlestranded, 22-ribonucleotide-average-sized, non-coding RNAs-modulate post-transcriptional gene expression by either repressing translation or degrading their target messenger RNAs (mRNAs) after binding to the mRNAs 3′-untranslated region 1,2 . In the brain and central nervous system (CNS) miRNAs appear to act as single-discrete single-stranded RNA molecules or may be encapsulated into plasma membrane-derived exosomes (EXs) or extracellular microvesicles 3 .
In the brain-including the neocortex, hippocampus, and the limbic system-miRNAs regulate several intracellular signaling pathways involved in synaptic homeostasis, neuronal bioenergetics activity, and protein/lipid homeostasis 4,5 . In general, miRNA expression patterns are complex and dynamic; for instance, the natural miRNA abundance is subject to alterations during neural development and differentiation of the human brain and in the CNS during aging [6][7][8] .
Emerging data encourage further analysis and characterization of specific miRNAs that may be either downregulated or upregulated aberrantly in the prodromal and preclinical phases of AD 7,8,[14][15][16] , where treatments with putative disease-modifying effect are more likely to be effective [17][18][19] .
We sought to examine the impact of primary biological factors-age, sex, and apolipoprotein E (APOE) ε4 alleleon plasma concentrations of the selected miRNA set. Subsequently, we performed an exploratory study, both cross-sectional and longitudinal, to investigate whether baseline concentrations of these six miRNAs are associated with established neuroimaging biomarkers of AD pathophysiology, namely (I) Aβ-positron emission tomography (Aβ-PET), at both global and regional level, and (II) 18 F-fluorodeoxyglucose-PET ( 18 F-FDG-PET) to assess neuronal metabolism.

Study participants and cognitive assessment
Sixty individuals with SMC were recruited from the standardized, large-scale, observational, monocentric, French academic university-based "INveStIGation of AlzHeimer's PredicTors in Subjective Memory Complainers" (INSIGHT-preAD) study 35 , that is part of the Alzheimer Precision Medicine Initiative (APMI) and its established Cohort Program (APMI-CP) 36 . Participants were enrolled at the Institute of Memory and AD (Institut de la Mémoire et de la Maladie d'Alzheimer, IM2A) at the Pitié-Salpêtrière University Hospital in Paris, France.
The main objective of the INSIGHT-preAD study is to explore the earliest preclinical stages of AD and their progression to incipient objective cognitive impairment, using comprehensive clinical parameters and multi-modal biomarkers.
In brief, the INSIGHT-preAD study includes 318 cognitively and physically normal Caucasian individuals, recruited from the community in the wider Paris area, France, aged 70 to 85, with SMC. The status of SMC was confirmed as follows: (I) participants gave an affirmative answer ("YES") to both questions: "Are you complaining about your memory?" and "Is it a regular complaint that has lasted for more than 6 months?"; (II) participants showed intact cognitive functions based on the Mini-Mental State Examination score (MMSE, ≥ 27), Clinical Dementia Rating scale (CDR = 0), and Free and Cued Selective Rating Test (FCSRT, total recall score ≥41). Aβ-PET imaging investigation was performed at the baseline visit, as a mandatory inclusion criterion. Thus, all individuals enrolled in the study have SMC and are stratified as either positive or negative for brain Aβ deposition. At baseline, demographic, clinical data, and APOE genotype (See Supplementary material) were collected in all participants. Exclusion criteria were a history of neurological or psychiatric diseases, including depressive disorders.
The study was conducted in accordance with the tenets of the Declaration of Helsinki of 1975 and approved by the local Institutional Review Board at the participating center (Ethical approval number: 2013-Fev-13150). All participants or their representatives gave written informed consent for the use of their clinical data for research purposes.
Blood sampling and collections of tube storage Ten (10) mL of venous blood were collected in one BD Vacutainer ® tube (lithium heparin), which was employed for subsequent analyses. Blood samples were taken in the morning, after a 12-hour fast, handled in a standardized way, and centrifuged for 15 minutes at 2000 G-force at 4°C. Per sample, plasma fraction was collected, homogenized, aliquoted into multiple 0.5 mL cryovial-sterilized tubes, and finally stored at -80°C within 2 hours from collection.
Data for plasma concentrations of the six miRNAs were collected at participants' enrollment ("baseline visit" or "M0").
Total RNA isolation from human plasma Plasma samples from age-and sex-matched SMC individuals (N = 60), dichotomized according to the Aβ-PET imaging status-either positive (N = 30) or negative (N = 30)-were selected for the analysis. Samples were thawed slowly on ice, and total RNA was isolated from 200 μL of plasma using a miRNeasy Serum/Plasma Advanced Kit (Qiagen, Germantown, MD) according to the manufacturer's instructions for liquid samples. To allow for normalization of sample-to-sample variation in RNA isolation, 3.5 μL of 1.6 × 10 8 copies/μL of synthetic Caenorhabditis elegans miRNA cel-miRNA-39 (Qiagen) was added to each sample right after the addition of lysis buffer (buffer RPL), during the first steps of the RNA isolation protocol. RNA was, then eluted with 20 μL of H 2 O according to the manufacturer's instructions.

Measurement of plasma miRNAs concentration
The samples were treated with heparinase during the reverse transcription (RT) step, as previously described elsewhere 37 , to overcome the inhibition from residual heparin to downstream real-time quantitative PCR (qRT-PCR) reactions. In brief, 0.25 μL of RNA was reversetranscribed using the TaqMan High-Capacity cDNA Reverse Transcription Kit and miRNA-specific RT stemloop primers (Thermo Fisher Scientific, Waltham, MA) in a 15-μL RT reaction containing 0.25 μL of the RNA extract, 0.15 μL of 100 mM dNTP, 1.0 μL of MultiScribe™ reverse transcriptase (50 U/μL), 1.5 μL of 10× RT buffer, 0.19 μL of RNase inhibitor (20 U/μL), 3.0 μL of 5× miRNA-specific stem-loop RT primer, 8.4 μL of RNasefree water and 0.5 μL of Bacteroides Heparinase I (12 U/ μL) (New England Biolabs, Ipswich, MA). The reaction mixture was first incubated at 37°C for 1 h for heparinase digestion, followed by RT reactions at 16°C for 30 min, at 42°C for 30 min, and at 85°C for 5 min, and then maintained at 4°C. The complementary DNA (cDNA) products were stored at −20°C until analysis.
For qRT-PCR, the RT products were diluted 3× prior to qRT-PCR. In all, 2 μL of the diluted cDNA product was employed as a template in a 20-μL reaction containing 1.0 μL of TaqMan miRNA Assay (Thermo Fisher Scientific), 7.67 μL of RNase-free water, and 10 μL of TaqMan™ Universal Master Mix II, with UNG (Thermo Fisher Scientific). The qRT-PCR was performed with Bio-Rad CFX96 real-time PCR system (Bio-Rad) at 95°C for 10 min, followed by 40 cycles of 95°C for 15 s and 60°C for 1 min (see Supplementary materials). Data were then analyzed with Bio-Rad CFX Manager (Bio-Rad) with the automatic Ct setting for assigning baseline and threshold for Ct determination. All quantities of the plasma miR-NAs are expressed as relative quantities based on the 2-ΔΔCt method in which the relative expression level of each specific miRNA was calculated using the 2-ΔΔCt method after normalization to the spiked cel-miRNA-39.
In essence, all significantly changed plasma miRNA levels reported in this study (out of every human plasma miRNA (N = 2650)) were quantified using both RNA sequencing methods and qRT-PCR, adding to the analytical strength and robustness for this type of investigation.

miRNA-based study design
A two-step pilot study was carried out on plasma samples from age-and sex-matched SMC individuals (N = 10) recruited from the INSIGHT-preAD study and dichotomized according to the Aβ-PET imaging status, either Aβ-PET positive (N = 5) or Aβ-PET negative (N = 5). Using RNA sequencing methods (LC Sciences, LLC, Houston, TX, US), significant intergroup differences were found in the plasma miRNA molecular signature [miRNA-125b, miRNA-146a, miRNA-15b, miRNA-148a, miRNA-26b, and miRNA-100]. Then, to explore this six-miRNA panel, the first step pilot study was expanded (N = 60) by recruiting further age-and sex-matched SMC individuals of the same cohort, dichotomized as Aβ-PET imaging positive (N = 30) or negative (N = 30). Outliers were removed before performing association and comparison analysis. The removal of the outliers was based on the 1.5 IQR (interquartile range = Q3-Q1) rule: data points outside of the Q1-1.5 IQR and Q3+1.5 IQR are considered as outliers. Data distribution among data groups may vary, including the number of outliers. However, in most cases, no more than two data points were removed in each group.

PET data acquisition and processing
Aβ-PET investigation was performed at the baseline visit ("M0")-as mandatory inclusion criterion-and at two-year follow-up ("M24").
Brain amyloid PET scans were acquired 50 minutes after injection of 370 MBq (10 mCi) of 18 F-Florbetapir. Brain 18 F-FDG scans were obtained 30 minutes after injection of 2 MBq/kg of 2-deoxy-2-( 18 F)fluoro-D-glucose ( 18 F-FDG). All acquisitions were performed in a single session on a Philips Gemini GXL scanner and consisted of 3 × 5 minutes frames with a voxel size of 2 × 2 × 2 mm 3 35 .
Reconstructed PET images are analyzed with a pipeline developed by the CATI team, a French neuroimaging platform (http://cati-neuroimaging.com), according to a method previously described 35,38 . For longitudinal analysis, the mean activity in supratentorial white matter (eroded with a radius of three), the pons and whole-cerebellum regions were used as a reference for individual voxel normalization in the partial volume effect corrected images, as previously suggested 39 .
Standard uptake value ratios (SUVR) were calculated for each of 12 bilateral cortical regions of interest (anterior and posterior cingulate, superior frontal, inferior parietal, middle temporal cortices, and precuneus), as well as the global average SUVR.
A threshold of 0.79 obtained from a cross-sectional pipeline previously published 35 was used in the present study to dichotomize individuals in Aβ-PET positive and Aβ-PET negative. Correlation analyses were performed through a longitudinal pipeline deployed through a stepwise process, previously published by our group 40 .
The same image-assessment pipeline was applied to measure brain glucose metabolism on 18 F-FDG-PET scans, in a separate session. Cortical metabolic indices were calculated through Automated Anatomical Labeling atlas-120 ROIs, from which we selected the bilateral anterior cingulate cortex, posterior cingulate cortex, inferior parietal lobe, precuneus, middle temporal cortex, and hippocampus, with the pons used as the reference region.

Statistical analysis
The analyses were conducted on a subset (N = 60) of the INSIGHT-preAD cohort with at least one plasma miRNA measure. First, we performed a comparison between Aβ-PET-positive and Aβ-PET-negative individuals seeking for the difference in plasma miRNA concentrations. Given the non-Gaussian distribution of miRNAs concentrations, we performed the Kruskal-Wallis one-way analysis of variance on ranks test.
Then, we investigated the effects of age, sex, and APOE ε4 allele carrier status on each plasma miRNA concentration at baseline, using linear models (LM).
Then, we tested the association between plasma miR-NAs concentration, at baseline, and brain Aβ accumulation, measured in terms of baseline and longitudinal SUVR, at both global and regional levels. Brain regions were a priori selected based on knowledge of the early stages of AD. To this aim, we used linear mixed models (LMM) with random intercept adjusted for age, sex, and APOE ε4 carrier status. To investigate the longitudinal association between plasma miRNAs concentration and brain Aβ accumulation, we tested the interaction effect between baseline miRNAs concentration and time (miR-NAs*time) on SUVR changes, over a 2-year follow-up ("M0" and "M24").
Then, using a similar approach, we tested the association between plasma baseline miRNAs concentration and both baseline and longitudinal brain hypometabolism, assessed via 18 F-FDG-PET, at the regional level.
Given the non-Gaussian distribution of miRNAs concentrations, we carried out permutation tests to investigate the statistical significance of the model coefficients. This method is more appropriate than parametric tests for small samples and when the sampling distribution is unknown 41,42 . Moreover, differently from ranksum tests, permutation tests allow for interpretation of the results as the difference of median between groups 43 . Finally, permutation tests do not require estimation of statistic values or degrees of freedom. Thus, only P values were reported. In the present study, 5000 permuted samplings were run to estimate P values.
The full set of statistical analyses was conducted using the R software, version 3.6.0. Permutation tests for LM and LMM were computed using the "pgirmess" and "predictmeans" libraries, respectively, both available at http://cran.r-project.org/web/packages. P values < 0.05 were considered significant in all statistical elaborations. Given the explorative character of the present study, no correction of P values for multiple comparisons was applied.

Results
Participant characteristics, Aβ-PET group comparisons for miRNA concentrations, and association with key biological factors Sociodemographic features, APOE ε4 allele frequencies, Aβ-PET groups, and plasma baseline concentrations of the six miRNAs of the individuals are reported in Table 1.
Among the miRNAs identified in the total sample of 60 individuals, only miRNA-15b was different between the two Aβ-PET subgroups, being significantly lower in the Aβ-PET-positive individuals than Aβ-PET-negative ones (see Table 2 and Fig. 1).
No association between plasma concentrations of the six miRNAs and age was found (P > 0.145). Plasma concentrations of all miRNAs did not significantly differ between males and females (P > 0.214).
Likewise, we did not observe a significant APOE ε4 allele effect on plasma baseline miRNAs concentrations (P > 0.083).

Effect of plasma baseline miRNAs concentrations on the rate of brain Aβ deposition
We investigated whether plasma baseline concentrations of each miRNA are associated with baseline (M0) and follow-up (M24) brain Aβ accumulation, at both the global and regional level.
At M0, we found a negative association between plasma miRNA-100 concentrations and Aβ-PET SUVR in the posterior cingulate cortex (left P = 0.046; right P = 0.049). A negative association was also observed between plasma miRNA-125b concentrations and Aβ-PET SUVR in the right anterior cingulate cortex (P = 0.047).
LMM also showed a significantly negative association of the miRNA-125b*time and miRNA-148a*time interactions with Aβ-PET SUVR in the right anterior cingulate cortex (P = 0.030 and P = 0.045, respectively), over a two-year follow-up (M0 and M24).
No plasma miRNA was associated with global Aβ-PET SUVR. We did not find any significant associations between regional Aβ-PET SUVR and plasma miRNA-15b, miRNA-26b, and miRNA-146a.

Effect of plasma baseline miRNAs concentrations on brain metabolism
We investigated whether plasma baseline concentrations of each miRNA were associated with baseline (M0) and follow-up (M24) of the regional 18 F-FDG-PET signal.
We found a positive effect of the miRNA-15b*time interaction in the left hippocampus (P = 0.021) and a negative miRNA-26b*time interaction in the left precuneus (P = 0.048).
No significant associations with 18 F-FDG-PET signal were found for plasma miRNA-100 and miRNA-146a.

Discussion
To our knowledge, this is the first study investigating the association of miRNA-based signatures in plasma with AD core neuroimaging biomarkers of the Aβ pathway and neuronal metabolism, in a cohort of cognitively normal individuals with SMC, a condition at increased risk for AD.
We observed no significant association between age, sex, APOE ε4 status, and individual miRNA abundance or speciation, suggesting that miRNAs may represent an independent variable in specific neurological disease states. In similar studies analyzing the influence of the confounding factors age and sex on miRNA profiles from peripheral blood, it was found that very few single miR-NAs remained significantly associated with age and sex  after adjustment using multiple testing parameters 44 . This lack of significant association may also be a reflection of (I) the observation that many miRNA biomarkers show discrepant results in independent investigations of the same disease, and (II) the remarkable heterogeneity of miRNA complexity and expression patterns in neurological disorders, such as those observed across the entire AD continuum 45 . We report significantly lower levels of miRNA-15b in Aβ-positive individuals than -negative ones.
Interestingly, plasma miRNA-15b, integrated into a panel of other miRNAs, was found specifically downregulated in AD patients compared with both normal controls 46,47 and Parkinson's disease patients 47 ; it also positively correlated with cerebrospinal fluid (CSF) Aβ 1-42 concentrations 47 . In another study, the profiling of exosomal miRNAs extracted from serum samples disclosed miRNA-15b downregulated expression in AD patients compared to healthy control individuals 48 .
Such evidence led to infer a potential link between miRNA-15b and protective dynamics counteracting the dysregulation of the Aβ pathway. This hypothesis is supported by studies conducted in SH-SY5Y/APPswe neuroblastoma cell lines that reported an association of miRNA-15b with: (I) inhibition of brain Aβ accumulation, by directly downregulating β-site amyloid precursor protein cleaving enzyme 1 (BACE1) protein expression or (I) attenuation of Aβ-induced secretion/expression of pro-inflammatory cytokines, by suppressing the proinflammatory nuclear factor kappa-light-chain-enhancer of activated B cells signaling pathway activation 49 . This experimentally inferred anti-Aβ effect alongside the statistical difference between our two Aβ-PET subgroups of individuals, suggested an association between miRNA-15b and regional Aβ-PET SUVRs that we did not find. Although there may be a neurobiological explanation, we cannot rule out a statistical Type II error related to the small sample size (see also below).
We also found a positive association between miRNA-15b concentrations and levels of neuronal metabolism ( 18 F-FDG-PET SUVR) in the hippocampus. Although an interpretation of this finding is rather challenging, it is worth noting that miRNA-15b is known for its regulatory activity of cortical development and its involvement in the expansion and differentiation of cortical neural progenitor cells 50 .
We showed associations of miRNA-125b and miRNA-148a concentrations with Aβ-PET SUVR and miRNA-15b with 18 F-FDG-PET signal in some early AD-related brain regions and the hippocampus.
In the present study, the negative baseline association between plasma miRNA-125b and miRNA-100 concentrations and Aβ-PET SUVR in the right anterior cingulate cortex and the bilateral posterior cingulate cortex indicate a potential modulatory role of these miRNAs on the Aβ pathway, tracked at the fibrils and plaques phase when investigated with the current Aβ-PET radiotracers. The negative association of miRNA-125b and miRNA-148a concentrations with baseline and follow-up Aβ-PET SUVR in the right anterior cingulate cortex strengthens such a hypothesis.
The negative associations between these three miRNAs (miRNA-125b, miRNA-100, and miRNA148a) and baseline and follow-up levels of Aβ-PET radiotracer uptake support a knowledge-based, albeit speculative, hypothesis that the miRNAs in question are linked to molecular mechanisms that inhibit Aβ deposition.  Table 1 for descriptive data). Plasma miR-15b concentration is reported in terms of relative quantity. PET positron emission tomography, Aβ amyloid beta, n.s. not significant, miRNA microRNA.
In an attempt to address our results from an overarching neurobiological standpoint, we outline that miRNA-125b is the miRNA more extensively investigated in the AD human brain and biofluids and experimental models of the disease as well. Studies conducted in primary mouse cortical neurons (MCN) and neuroblastoma Neuro2a (N2a) cells show that miRNA-125b binds to the BACE1 protein with an inhibitory effect on its expression levels and mitigates Aβ-induced neurotoxicity 51 . In particular, BACE1 protein acted as a target of miRNA-125b and was negatively modulated by this miRNA in both MCN and N2a cells. BACE1 restoration inhibited the effect of miRNA-125b on Aβ-induced neurotoxicity, thus indicating that miRNA-125b might repress AD development by reducing BACE1 expression. Moreover, the authors showed that toxic downstream effects of Aβ species, such as apoptotic rate, promotion of caspase-3 activity (crucial for neuronal apoptosis), and oxidative stress-were all decreased. The same study also reported that miRNA-125b and BACE1 mRNA serum expression levels were respectively reduced and increased in patients with sporadic AD compared to age-matched healthy individuals. A negative correlation between miRNA-125b and BACE1 mRNA serum expression levels was found in AD patients 51 .
CSF and blood (serum)-based studies reported downregulation of miRNA-125b in AD patients compared with a cohort of non-inflammatory and inflammatory neurological disease controls 52 and to healthy controls 53 . The same result was reported in two independent studies conducted in either CSF or plasma 54 .
By contrast, our miRNA-125b-related result seems to counteract a previous post-mortem study reporting that miRNA-125b was significantly and distinctively upregulated in AD temporal lobe neocortex, potentially contributing to the neurodegenerative process of the disease 55 . Another study reported that miRNA-125b is also overexpressed in the hippocampus in AD late stages (Braak stages V, VI) 56 .
The studies mentioned above support the negative (five out of eight regions, with only one region proving positively associated and counteracting this trend of associations) association we found between miRNA-125b and 18 F-FDG-PET indexes in multiple brain regions involved in early AD. This finding is in line with experimental studies that showed an association between miRNA-125b and synaptic toxicity in cultured hippocampal neurons 57 as well as between miRNA-125b overexpression and impairment of synaptic remodeling and transmission coupled with inhibition of the long-term potentiation (LTP) activation 58 .
Moreover, miRNA-125b, one of the most highly expressed microRNAs in the human brain and retina, was associated with kinase/phosphatase activity inbalance, increased tau phosphorylation and its downstream perturbation of the synaptic activity 59 .
In the present study, the absence of converters from SMC to mild cognitive impairment (MCI) (or even overt dementia) does not allow drawing any meaningful biological or clinical conclusion on the role of miRNA-125b in AD; however, our PET-based results coupled with the existing literature suggest that miRNA125b may be linked to Aβ-independent neurotoxic pathways. The limitations of the present exploratory study will need to be overcome in future investigations with longer followup and higher rates of conversion to MCI and AD dementia to untangle the role, either mono-dimensional or dualistic, of miRNA-125b.
There is no extensive experimental evidence to explain the unprecedented association we here report for miRNA-148a and miRNA-100 with biomarkers charting the Aβ pathway and neuronal metabolism. However, these miR-NAs have been previously investigated in AD, with the former upregulated in AD patients 60 , and miRNA-100 found lower in AD versus non-demented controls 61 .
Neurobiological studies pointed out miRNA-148a as upregulated in the hippocampus of 3xTgAD mice versus their age-matched controls 62 , and it was highly concentrated in synapses 63 and downregulated the expression of Ca 2+ /calmodulin-dependent protein kinase IIα 64 , a key protein for Calcium homeostasis. Calcium dysregulation has been demonstrated to potentially occur upstream to brain proteinopathies with Aβ monomers production and tau hyperphosphorylation that can be triggered by a sustained increase of cytosolic Calcium concentrations over basal levels 65 . From a functional standpoint, miRNA-148a upregulation was associated with impairment of LTP that underlies memory dysfunction in AD 64 .
Regarding the positive association found between miRNA-148a and 18 F-FDG-PET SUVR, at both baseline and follow-up, the interpretation of this data is rather challenging given the lack of experimental evidence about this miRNA and neuronal functions.

Limitations
The study presents some caveats that need to be addressed. First, the restricted sample size of the first step of the pilot study (N = 10) potentially limited the number of miRNAs dysregulated that were detected between groups. Moreover, the relatively small sample size of the second step pilot study (N = 60) limited the generalizability of our conclusions. The absence of converters to MCI or even dementia stage constrains the clinical conclusions that can be drawn. Future studies in larger cohorts of SMC individuals, with extensive follow-up, are needed to corroborate or disconfirm the findings of the present pilot data.
We performed a priori investigation on selected brain regions rather than executing a whole-brain exploratory strategy. Although the latter approach might have contributed to untangle patterns of brain Aβ accumulation and synaptic activity, we could not run it owing to statistical reasons (i.e., statistical power).
Finally, our investigation was exclusively performed on a sample of Caucasian individuals; hence, replicating our findings in different ethnic populations would be remarkable.
To increase the clinical meaningfulness of the present exploratory study results, we intend to carry out a subsequent study in a larger population of INSIGHT-preAD cognitively healthy individuals, possibly employing an independent validation cohort, and including (i) bloodbased biomarkers charting AD pathophysiological alterations and (ii) longer follow-up of neuroimaging and cognitive measures. We set out to explore whether plasma miRNA can predict AD clinical-biological trajectories; thus, representing a candidate non-invasive and accessible screening and prognostic tool. If our results were corroborated, plasma miRNA could be integrated into a bloodbased biomarker matrix for precision medicine-oriented large-scale investigation of cognitively healthy individuals at increased risk of developing rapid AD pathophysiological and cognitive impairment, a highly suitable population for disease-modifying therapeutic approaches when these will be available.

Conclusions
This is the first exploratory study reporting significant associations between plasma concentrations of a signature of miRNAs we identified and AD core neuroimaging biomarkers of the Aβ pathway and neuronal metabolism in a cohort of SMC individuals, a condition associated with increased risk of AD. In particular, our study provides preliminary evidence in vivo about the potential role of miRNAs -125b and -15b, as candidate miRNA biomarker of AD pathophysiology.
In summary, our findings are consistent with the endevour of generating insights about the roles that circulating miRNAs may play for AD pathophysiology. Results turning out of this research line may inform precision-medicine oriented pharmacological trials that take into account the entire biological profile of the single patient. Such an approach could be then generalized to the broad spectrum of neurodegenerative diseases by developing pathway-based therapies in biologically ideal individuals, in line with the precision medicine paradigm shift.