Background
Amyloid β (Aβ) and tau pathologies are classic characteristic features of Alzheimer’s disease (AD), and they are widely used as diagnostic biomarkers [
1]. Aβ and tau burden in the brain can be identified with high accuracy from cerebrospinal fluid (CSF) testing [
2] and positron emission tomography (PET) imaging [
3,
4]. However, the high cost and low availability of PET scans hamper the feasibility of their use in clinical diagnostic practice and clinical trials. Aβ and tau in samples of CSF obtained from patients has been shown to diagnose AD with excellent accuracy [
5]. Multiple studies have suggested that the combined measurements of phosphorylated-tau (P-tau) and Aβ42 in the CSF can inform a more accurate diagnosis than either test alone; this improved diagnostic accuracy is likely due to the reduced impact of preanalytical and analytical confounders [
6‐
8]. Further supporting the CSF ratio of P-tau/Aβ42 as a reliable diagnostic biomarker for AD, several studies have reported similar threshold levels, in the range of 0.09–0.14 [
7,
9‐
11]. However, the relatively invasive nature of CSF collection restricts its use as a screening tool in the elderly population. Hence, there is an unmet need for a minimally invasive, widely available, and cost-effective method of measuring biomarkers for the early detection of AD in the general population.
By measuring a panel of microRNAs (miRNAs) in the blood, the current study proposes a simple, antibody-independent method of predicting the P-tau/Aβ42 ratio in the CSF. miRNAs are short non-coding RNAs of approximately 20–25 nucleotides in length that bind to complementary sites on the 3′ untranslated region (UTR) of their mRNA targets, curbing their expression [
12]. Changes in miRNA expression may induce translational abnormalities, resulting in the alteration of corresponding protein levels. An increasing number of studies have demonstrated a relationship between miRNAs and AD; by targeting the expression of amyloid precursor protein (APP) or beta-site APP cleaving enzyme 1 (BACE1) [
13], miRNAs can directly affect potential pathogenic pathways and thus alter the risk and/or progression of AD. A panel of 12 miRNAs can reportedly diagnose AD with high performance [
14], indicating the combination of miRNA panels as a promising biomarker for AD. However, a recent review paper listed 48 studies on circulating microRNAs as potential biomarkers for AD, which showed inconsistent data [
15]. The first potential reason of the inconsistencies among these studies may be the small sample sizes of these studies. The sample size of these studies ranged from 6 to 287 (AD patients), and 29 studies (over 60.4% of 48 studies in total) included a sample size of < 30 AD patients. The too-small-sample-size studies may produce bias in the results. The second reason for the inconsistencies would be that most studies did not use CSF or PET biomarkers to recruit AD patients. In this study, the strict inclusion criteria involving CSF biomarkers and a large sample size were recruited, which may guarantee the potential clinical application of positive findings.
In addition, AD and other types of dementia, such as vascular dementia (VaD), Parkinson disease dementia (PDD), behavioral variant frontotemporal dementia (bvFTD), and dementia with Lewy body (DLB), may have overlapping clinical manifestations, pathology, and biomarkers, often resulting in difficulties in clinical diagnosis [
16]. Whether miRNAs can differentiate AD from other forms of dementia has been addressed by few studies. Given the crucial role of miRNAs in the expression of genes that are key to AD pathology, their relative stability, tissue enrichment, and amenability to quantitative measurement [
15], we speculated that measuring single or multiple miRNAs may reflect the concentration of Aβ and tau in the brains of AD patients. Therefore, this study aimed to evaluate whether the levels of blood miRNAs (1) predict the P-tau/Aβ42 ratio in the CSF, (2) can be used to differentiate patients with AD from cognitively normal controls, and (3) can effectively discriminate AD from VaD, PDD, bvFTD, and DLB.
Discussion
The present study identified an association between a panel of blood miRNAs and the ratio of P-tau/Aβ42 in the CSF of patients with AD, suggesting miRNAs as a promising tool for predicting the Aβ42 and P-tau levels in patients with AD.
Biomarkers have played an important role in the diagnosis [
1] and research [
17] of AD. Because of its minimal invasiveness and relatively low cost, the use of peripheral blood to diagnose AD has garnered increasing attention. The attendant surge in research has revealed a series of promising markers in the blood, including Aβ42 [
34], the neurofilament light protein (NFL) [
35], P-tau181 and 217 [
36], exosomal Aβ42, T-tau, P-tau, synaptic proteins, and inflammatory factors [
10,
11,
37]. Despite their high diagnostic efficacy, this method is subject to limitations. Requiring advanced skill and specialized equipment, the collection and measurement of Aβ42 from the blood by immunoprecipitation coupled with mass spectrometry is cost-prohibitive. Moreover, NFL is not a specific biomarker for AD; aberrant NFL concentrations may indicate other diseases causing axonal damage, such as multiple sclerosis (MS) [
38], frontotemporal dementia (FTD) [
39], and amyotrophic lateral sclerosis (ALS) [
40]. While blood P-tau can be easily measured, it requires a specialized testing system that may require further development before its cost can allow for extensive, wide-spread use [
41]. The screening of biomarkers from exosomes in the blood may be excessively expensive, as it requires the collection and enrichment of neuron-derived exosomes through a series of experiments, including immunoprecipitation and ELISA.
By contrast, the analysis of miRNAs in blood is an antibody-independent and easily implemented method for differentiating patients with AD from their cognitively normal counterparts, as well as patients with other forms of dementia. By only requiring the widely used technique of qPCR to quantify a panel of serum miRNAs, our technique can predict the P-tau/Aβ42 ratio—a well-known AD biomarker—in the CSF. To the best of our knowledge, this study is the first to provide support for an association between miRNAs in the blood with P-tau/Aβ42 in CSF and is a promising application to screen for AD in older populations at relatively little cost and with minimal invasiveness.
Recent studies have increasingly implicated miRNAs in AD pathology; miRNAs regulate the expression of APP [
42‐
45] and proteins involved in APP metabolism, such as α-secretase, ADAM10 [
46,
47], β-secretase, and BACE1 [
48,
49]. miRNAs also play an important role in Aβ clearance, e.g., miRNAs can downregulate ApoE lipidation [
50] and TREM2 levels [
51] and impair Aβ metabolism in the brain. Moreover, miRNA levels are related to the expression and hyperphosphorylation of tau in the brain [
52‐
54] and are involved in other AD-associated mechanisms, such as aberrant mitochondrial function [
55‐
57], autophagy [
58,
59], mitophagy [
60,
61], neurotransmitter release and clearance [
62,
63], and synaptic plasticity [
64].
Due to their important roles in the pathology of AD, miRNAs can act as biomarkers of the disease [
65]. miRNAs have been used as biomarkers for a range of diseases, such as cancer [
66,
67], cardiovascular disease [
68,
69], and diabetes [
70,
71]. In agreement with the observations of dysregulation of miRNAs in the CSF of patients with AD [
72], alterations of miRNAs in the peripheral blood have shown potential as promising candidate biomarkers of AD. The combination of several miRNAs was able to discriminate the CSF of patients with AD from that of controls with sufficient accuracy [
73]. A recent literature review showed that, among 137 miRNAs found to be abnormally expressed in the blood of patients with AD, 36 had been replicated independently in more than one study [
74]. This finding provides evidence in support of the use of miRNAs as diagnostic biomarkers. Moreover, a signature of 12 miRNAs in the blood could not only inform the discrimination between AD patients and controls but also between patients with AD and those with other neurological disorders, such as Parkinson’s disease and schizophrenia [
14]. While our findings confirm the utility of miRNAs as biomarkers for AD, our study further suggests that miRNAs could reflect the P-tau/Aβ42 ratio in the CSF, an established AD biomarker. We attribute this association to the important roles of miRNAs in the regulation of AD pathways in the brain. We further compared miRNA levels between AD and VaD, PDD, bvFTD, and DLB. Although all these degenerative diseases have some similar clinical manifestations such as cognitive impairment, AD has its unique pathological process, which may be the reason why the changes of these seven miRNAs are AD-specific and differentiates AD from other neurodegenerative diseases. However, our results concerning the upregulation or downregulation of a single miRNA were inconsistent with the observations of other studies: while miR-17 was reported to be significantly altered in the blood of AD patients [
75], our own study could not confirm this. This discrepancy might suggest that miRNA expression may vary according to ethnicities. Further multi-center studies are needed to evaluate genetic differences in miRNA expression between different ethnic populations.
This study is limited by its cross-sectional design. Although we confirmed that a panel of seven miRNAs could be applied as diagnostic biomarkers of AD, longitudinal designs would be better suited for the evaluation of the performance of these biomarkers. Hence, longitudinal studies investigating the relationship between the levels of biomarkers and the decline in cognitive functions of patients are warranted. This study was further limited by its not having considered patients with mild cognitive impairment that progressed to either AD or stable amnestic mild cognitive impairment. The application of our method to the prediction of the progression from prodromal to probable AD is thus diminished. Finally, measuring miRNAs with qPCR is a relative quantification approach that cannot indicate absolute levels of miRNAs in the blood, limiting the comparisons of the absolute levels of miRNA between our study and others.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.