Exploring small non-coding RNAs as blood-based biomarkers to predict Alzheimer’s disease

Background Alzheimer’s disease (AD) diagnosis relies on clinical symptoms complemented with biological biomarkers, the Amyloid Tau Neurodegeneration (ATN) framework. Small non-coding RNA (sncRNA) in the blood have emerged as potential predictors of AD. We identified sncRNA signatures specific to ATN and AD, and evaluated both their contribution to improving AD conversion prediction beyond ATN alone. Methods This nested case–control study was conducted within the ACE cohort and included MCI patients matched by sex. Patients free of type 2 diabetes underwent cerebrospinal fluid (CSF) and plasma collection and were followed-up for a median of 2.45-years. Plasma sncRNAs were profiled using small RNA-sequencing. Conditional logistic and Cox regression analyses with elastic net penalties were performed to identify sncRNA signatures for A+(T|N)+ and AD. Weighted scores were computed using cross-validation, and the association of these scores with AD risk was assessed using multivariable Cox regression models. Gene ontology (GO) and Kyoto encyclopaedia of genes and genomes (KEGG) enrichment analysis of the identified signatures were performed. Results The study sample consisted of 192 patients, including 96 A+(T|N)+ and 96 A-T-N- patients. We constructed a classification model based on a 6-miRNAs signature for ATN. The model could classify MCI patients into A-T-N- and A+(T|N)+ groups with an area under the curve of 0.7335 (95% CI, 0.7327 to 0.7342). However, the addition of the model to conventional risk factors did not improve the prediction of AD beyond the conventional model plus ATN status (C-statistic: 0.805 [95% CI, 0.758 to 0.852] compared to 0.829 [95% CI, 0.786, 0.872]). The AD-related 15-sncRNAs signature exhibited better predictive performance than the conventional model plus ATN status (C-statistic: 0.849 [95% CI, 0.808 to 0.890]). When ATN was included in this model, the prediction further improved to 0.875 (95% CI, 0.840 to 0.910). The miRNA-target interaction network and functional analysis, including GO and KEGG pathway enrichment analysis, suggested that the miRNAs in both signatures are involved in neuronal pathways associated with AD. Conclusions The AD-related sncRNA signature holds promise in predicting AD conversion, providing insights into early AD development and potential targets for prevention. Supplementary Information The online version contains supplementary material available at 10.1186/s13578-023-01190-5.

Total RNA was isolated from 400 µL of plasma using the Plasma/Serum RNA Purification Midi Kit (Norgen Biotek, Canada) and quantified with the Qubit RNA HS Assay Kit (Life Technologies, USA) on a Qubit fluorometer (Life Technologies, USA).Since the concentration of RNA was below the lower standard, the entire volume was concentrated in vacuo using the Savant SPD2010 Speedvac Concentrator (ThermoScientific, USA) and resuspended in 5 µL of DNase/RNasefree water.SncRNAs were analyzed using Illumina TruSeq SncRNA Library Prep Kit (Illumina, USA), following the manufacturer's instructions.Briefly, RA5 and RA3 RNA oligonucleotides were ligated to 5' and 3' ends of RNA, respectively.Adapterligated RNA was reverse-transcribed using an RT Primer and the resulting cDNA was amplified in a 15-cycle PCR that used RP1 and indexed RP1 primers.A gel purification step was used to isolate PCR products of 145-160 bp through a 5% Mini-Protean TBE Gel (Biorad, USA).The integrity of the generated sncRNA sequencing library was confirmed using High Sensitivity D1000 Screen Tape Assay on 4200 TapeStation System (Agilent, USA), and it was quantified with the Qubit 1X dsDNA HS Assay Kit (Life Technologies, USA).The sncRNA expression profiling was generated from the NextSeq 2000 Sequencing System (Illumina, USA).Leftover RNA and cDNA were stored at -80ºC.
The Illumina DRAGEN Bio-IT Platform was used for the generation of raw, demultiplexed sequencing data.Quality control was performed using the FastQC (v0.11.9) software.Reads with a quality score of 20 or higher were filtered using Trim Galore tool (v0.6.7), which implements the Cutadapt tool for adapter trimming.
The mapping and quantification of sncRNAs were done in two different protocols: miRNA and other sncRNAs.For the identification of miRNAs, reads were mapped to the human genome(1) and quantified using mature and hairpin miRNA database references(2) using the miRDeep2 tool (v2.0.1.2).Duplicates were removed.For the identification of other sncRNAs, a BLAST (v.2.5.0) against the miRNA database was performed (2), kept the non-miRNA reads and a new BLAST against the sncRNA database(3) was performed.B2M is part of the MHC I complex, which exhibits increased expression in samples from AD patients, and impairs neuronal plasticity, neurite growth, and neurite regeneration.
It is a membrane marker of senescence and blood biomarker of AD.The increase in B2M is related to a decline in cognitive scores, to the pathological changes of Aβ (low CSF Aβ1-42), but not to neurodegeneration or tau pathology.
Compared with young mice, old mice experience elevated B2M levels in their hippocampus and plasma.In another study, HSP90AA1 was decreased in the blood from MCI and AD and was negatively correlated with α and β secretase activities. (

eFigure 2 |
Heatmap of sncRNAs correlations.Spearman's rank correlation coefficient is shown for each correlation.Only significant correlations are shown (P value < .05). 0

2 ↓ 1 ↓ 1 ↓ 2 ↑ATXN1 1 ↑ 1 ↑
miR-224-5p ↑ hsa-miR-221-5pInteraction between GAPDH and AD proteins through oxidative stress in brain.Specimens of AD, including Aβ, Aβ precursor protein and Tau, leads to impairment of the GAPDH glycolytic function in AD and thus in apoptosis.Proapoptotic factor abundant in the blood of AD patients.(28-31)SEC24C SEC24 homolog C, COPII coat complex component ↑ hsa-miR-877-5p ↓ hsa-miR-224-5p ↑ hsa-miR-221-5p The family of SEC24p oversee vesicle trafficking, including shaping vesicle, cargo selection and concentration, of subunits of the γ-secretase (involved in Aβ production).SEC24C transports the γ-subunit nicastrin.hsa-miR-224-5p ↓ hsa-miR-382-5p In spinal cord injury, ARIH2 overexpression promotes neurotrophic growth factor expression, brain-derived neurotrophic factor expression, astrocyte and neuronal viability, neuroprotective cytokine pleiotrophin, neurite outgrowth, and suppressed astrocyte apoptosis.Knockdown of ARIH2-derived Triad1 protein has the opposite effects.hsa-miR-625-5p ↑ hsa-miR-221-5p Downregulated ELAVL1 may reduce the abundance and type of mRNA in AD patients through modulation of mRNA stability and promotion of variable cleavage of precursor RNA.miR-548d-5p ↓ hsa-miR-769-5p ↓ hsa-miR-224-5p TNRC6A promotes dendritic growth, both in the hippocampus and cerebellum, by regulating global somatodendritic translation and actin cytoskeletal dynamics of developing neurons.Perturbation of this protein results in reduced dendritic growth of cultured hippocampal neurons.hsa-miR-625-5p ↓ hsa-miR-224-5p This oxygen-dependent transcription factor is the major regulator of EPO, which has neuroprotective and proangiogenesis effects, and triggers responses to hypoxia including metabolic adaptation and angiogenesis.hsa-miR-548d-5p ↑ hsa-miR-548k ↓ hsa-miR-625-5p It cleaves mRNA to regulate gene expression.Astrocytes exposed to Aβ1-42 oligomers show loss of Ago2 phosphorylation, which leads to uncontrolled production of pro-inflammatory cytokines and neuroinflammation, leading to loss of neurons.BACE1 and APP protein levels are significantly elevated following AGO2 silencing in cultured human brain cells.may help to shuttle ATXN1L into the nucleus where they interact in a large protein complex.Therefore, it is related to Aβ pathology mediated by BACE1 expression.Also related to progressive non-fluent aphasia.miR-221-5p ↓ hsa-miR-382-5p Involved in circadian rhythms during early AD development.The expression of Csnk1E was higher in AD mice compared to control mice after light induction.hsa-miR-877-5p ↓ hsa-miR-625-5p ↓ hsa-miR-224-5p ↓ hsa-miR-454-5p ↓ hsa-miR-382-5p Knockout of Atxn1 increases BACE1 levels and enhances amyloidogenic cleavage of APP, exacerbating Aβ deposition and gliosis in AD mouse models, and impaired hippocampal neurogenesis and olfactory axonal targeting.hsa-miR-548d-5p ↑ hsa-miR-877-5p Promotes hippocampal LTP.In AD patients and in a mouse model of AD it is decreased, inhibiting septal cholinergic neurons projecting to the hippocampus.(45) CRKL CRK like proto-oncogene, adaptor protein ↑ hsa-miR-877-5p ↓ hsa-miR-769-5p Some of the neurotoxic effects of Aβ peptides may be mediated via the activation of proteins belonging to the Abl family, that regulate actin cytoskeleton structure through CRKL and phosphorylate microtubule-associated Tau protein, thus weakening synaptic transmission.(46) NF2 Moesin-ezrin-radixin like (MERLIN) tumor suppressor ↓ hsa-miR-769-5p ↑ hsa-miR-221-5p Loss of function of the gene lead to the development of schwannomas, meningiomas and juvenile cataracts.Akt signaling pathway is anti-apoptotic, and AKT1 inhibits Akt phosphorylation (activation), which in turn upregulates the activity of GSK3β, a kinase implicated in the pathogenesis of AD.ROS-mediated oxidative modification of Akt1 leads to reduced synaptic Akt1-mTOR signaling which contributes to synaptic dysfunction in AD.It is rescued by overexpression of Akt1 in APP/PS1 mice.miR-625-5p ↓ hsa-miR-224-5p Regulates the stabilization and function of p-tau in AD.Inhibition of Hsp90 in AD mouse models reduced Aβ toxicity and normalizes synaptic function.
Characteristics of the A+(T|N)+ study participants according to their A, T and N profile.
Sensitivity analysis using logistic regression analysis to examine the individual associations between sncRNAs expression and A+(T|N)+.Sensitivity analysis of sncRNAs associated with A+(T|N)+ in a logistic regression analysis.SncRNA are ranked from the highest to the lowest elastic net positive and negative regression coefficients for A+(T|N)+.Cox regression analysis examining the individual associations between sncRNAs expression and risk of progression from MCI to AD. Genes regulated by the ATN-related sncRNA signature.Antioxidant enzyme that maintains the control of ROS.SOD2 polymorphisms have been associated with increased susceptibility to oxidative stress.The relationship between SOD2 rs4880 polymorphism and AD have yielded inconsistent results, but in combination with APOε4 allele carriage it increases the risk for MCI and AD.Genes regulated by the AD-related sncRNA signature.