Researchers say this breakthrough could lead to more effective prevention strategies and future therapies.
From a recent study published in Alzheimer’s and dementiaresearchers have identified the serum micro-ribonucleic acid (miRNA) signature in Alzheimer’s disease (AD). They also examined miRNAs that could predict the transition from mild cognitive impairment (MCI) to Alzheimer’s disease (AD).
Background
Alzheimer’s disease is a neurodegenerative disease characterized by progressive cognitive decline. The identification of AD in advanced stages leads to poor treatment outcomes. Thus, new diagnostic techniques are needed to identify individuals in early-stage MCI (EMCI) or late-stage MCI (LMCI) and predict their conversion to AD.
Current diagnostic tools are invasive and expensive. MicroRNAs are short, non-coding RNAs that control system-level proteostasis. They are potential minimally invasive and inexpensive biomarkers for AD. MicroRNAs can affect multiple mRNA targets, act paracrine, and participate in interorgan communication. Furthermore, these molecules are highly stable in cell-free environments and resistant to thaw-freeze cycles, making them logistically desirable in clinical settings.
About the study
In the current study, researchers examined the miRNA signatures of early MCI, late MCI, and AD and determined whether the signatures correlate with clinical disease state using established AD biomarkers.
Researchers obtained serum samples from participants in the AD Neuroimaging Initiative (ADNI) to analyze miRNA expression. Small RNA sequencing analyzed serum samples from 803 of the 847 ADNI participants, excluding poor quality samples. Of the 803 participants, 272, 217, 149, and 165 belonged to the EMCI, LMCI, AD, and control groups, respectively. Linear mixed regressions analyzed miRNAs expressed in 95% or more samples with at least ten measurements in each sample.
Machine learning (ML) models tested the miRNAs with significant effect sizes in the regression analysis. The ML data includes the training and testing datasets (fraction, 0.3). Researchers excluded miRNAs without predictive power, that is, those with area under the receiver-operating characteristic curve (AUC) values ≤0.5. For each condition, selecting the top five miRNAs resulted in 73 candidates. The ML method derived all possible combinations of up to three miRNAs for these candidates.
The researchers compared the performance of serum miRNA with ADAScog13 levels and invasive biomarkers of AD in the cerebrospinal fluid (CSF). In addition to amyloid beta and phosphorylated tau protein levels in CSF, they compared Mini-Mental State Examination (MMSE) scores. They also analyzed follow-up phenotypic data provided by participants for 144 months after the blood draw. Gene Ontology (GO) analysis examined the biological pathways controlled by the miRNAs identified in the current study.
Results
The serum miRNA signature for AD included miR.98.5p, miR.142.3p, and miR.9985 (AUC, 0.7). Serum miR.369.3p, miR.590.3p, and miR.9985 identified EMCI (AUC, 0.7). The three miRNAs that showed the best performance to identify LMCI were miR.22.5p, miR.1306, and miR.4429 (AUC, 0.7). In terms of AD prediction, Abeta levels and Abeta/tTAU or Abeta/pTAU181 ratios in cerebrospinal fluid and MMSE assays performed better than the serum miRNA signature.
Serum miRs for MECI performed better than CSF biomarkers (AUC of 0.5 to 0.6) and the MMSE assay (AUC, 0.6). To predict LMCI, the miR signature performed similarly to CSF biomarkers (AUC, 0.7) and the MMSE (AUC, 0.7). To predict conversion of EMCI to AD, miR.18a.5p, miR.26b.5p, and miR.125b.5p (AUC, 0.7) performed better than CSF biomarkers (AUC, 0.6). In identifying LMCI to AD converters, serum miR.142.3p, miR.338.3p, and miR.584.5p (AUC, 0.8) outperformed cerebrospinal fluid biomarkers.
ADAScog13 estimated the conversion of early-stage MCI to Alzheimer’s disease (AUC, 0.6) and LMCI-AD (AUC, 0.7) with lower accuracy than the miRNA signatures. Using serum miR.532.3p and miR.1306.3p expression improved the accuracy of predicting the conversion of EMCI and LMCI to AD. Combining the miRNAs for EMCI prediction with MMSE data moderately improved the accuracy of identifying LMCI patients (0.75 from 0.71).
The miRNA profiles of EMCI, LMCI, and AD indicate distinct molecular mechanisms. MiRNAs representing EMCI showed specific associations with oxidative phosphorylation and ferroptosis. The findings are consistent with previous research showing that disrupted energy and iron metabolism occurs early in AD pathophysiology. In contrast, only in the late phase of the MCI signature of miRNAs did researchers find mechanisms indicating interleukin-17 (IL-17) signaling and vascular damage, which are associated with the development of AD.
Conclusion
The study showed that serum miRNA signatures can be used as biomarkers for Alzheimer’s disease and predict the transition from MCI to AD. Combining these signatures with neuropsychological tests such as the MMSE can increase the accuracy of AD predictions.
The findings are of general importance because using serum miRNAs to characterize the at-risk segment of the aging population could reduce invasive and costly testing. Future research should refine and confirm these features and integrate them with cognitive screening measures.