Mild Cognitive Impairment (MCI), currently termed as mild neurocognitive disorder by the Diagnostic and Statistical Manual of Mental Disorders-5 (DSM-5), is defined as a decline in cognitive functioning. MCI can be classified based on whether memory is impaired or not, as amnestic, and non-amnestic subtypes, respectively. Moreover, depending on the number of affected domains, MCI can be categorized as single or multiple domains. Other relevant characteristic is that MCI does not interfere in the daily activities (Sachs-Ericsson and Blazer, 2015). Although MCI is not a disease, it has garnered significant interest in recent years because it falls between normal aging and the early stages of dementia (Morris, 2005). It has been reported that the transition from MCI to dementia occurs at an annual rate of 12% in the general population, reaching up to 20% in higher-risk groups (Plassman et al., 2008; Petersen, 2006). This suggests that early identification and timely treatment of MCI may delay the onset of dementia. However, MCI is clinically heterogeneous, which has hindered efforts to identify and monitor it (Sabbagh et al., 2020).

The MCI identification is established considering previous cognitive functioning before memory complain onset, educational level, lifestyle, and ethnic differences (Kouzuki et al., 2022). Currently, the assessment of MCI relies on the sensitivity and specificity of cognitive and functional tests, such as the Mini-Mental Status Examination (MMSE), the Montreal Cognitive Assessment (MoCA), among others (Jia et al., 2021). However, despite their general utility, diagnostic methods for MCI are constantly evolving and not well standardized, which may lead to variations in results and a lack of consistency in the early and accurate diagnosis of MCI (Chen et al., 2021; Lee, 2023; López and Altuna, 2023).

Although molecular markers are not currently included in the diagnostic criteria (Dunne et al., 2021), their potential in identifying and assessing the risk of progression from MCI to dementia has captured the attention of medical research. Indeed, various biomarkers in cerebrospinal fluid (CSF), such as Aβ42 and tau protein, have been explored with the purpose of using them for differential diagnosis between MCI, Alzheimer’s disease (AD), and cognitively normal subjects (Dunne et al., 2021; Papaliagkas et al., 2023). However, assessing biomarkers in CSF involves an invasive procedure, thus there is an urgent need to identify more accessible and less invasive molecular markers present in peripheral fluids that allow for timely diagnosis of MCI (Álvarez-Sánchez et al., 2023; Hansson et al., 2010).

In this context, considerable attention has been directed toward microRNAs (miRNAs) as potential molecular markers for MCI. miRNAs are non-coding RNA molecules that play a crucial role in regulating gene expression (Ranganathan and Sivasankar, 2014; Ratti et al., 2020). These small molecules can repress specific genes by binding complementarily to the 3′ untranslated region (3’UTR) of mRNA, leading to either degradation or inhibition of translation (McNeill and Van Vactor, 2012). Interestingly, miRNAs have also been found to interact with other regions of mRNAs, including the 5’ UTR, the coding sequence, and gene promoters, where they can promote protein translation (Groth et al., 2007; Ørom et al., 2008). This versatility underscores their significance as regulators of cellular activity in both pathological and non-pathological conditions (O’Brien et al., 2018; Ranganathan and Sivasankar, 2014).

Recent studies have indicated that alterations in the expression of specific miRNAs may be linked to the regulation of biological processes involved in neurodegenerative diseases (Li et al., 2023; Mu et al., 2024). Furthermore, various investigations have shown that circulating miRNAs are present in peripheral fluids such as blood, serum, plasma, and saliva, as well as in extracellular vesicles like exosomes. These miRNAs can be easily detected and, due to their high stability, reflect subtle changes that occur during the early stages of neurodegeneration (Kamal and Shahidan, 2020; Kumar and Reddy, 2016).

To date, various studies have evaluated the expression of miRNAs in MCI. However, challenges exist in identifying reliable molecular markers mainly due to the lack of multi-platform standardization, the continuous discovery of new miRNAs, and the small sample size (Condrat et al., 2020). Therefore, integrative approaches such as meta-analysis, which enable the combination of data from multiple studies to identify the most relevant miRNAs, have emerged as a promising option. These methods not only contribute to increasing predictive capacity but also help overcome potential contradictions present in individual research (Roointan et al., 2023; Yuen et al., 2021).

Considering the above, our aim was to identify commonly dysregulated miRNAs in MCI and characterize their biological significance. To achieve this, we conducted a meta-analysis encompassing previously published studies on miRNA expression profiles in samples of plasma, serum, and extracellular vesicles. Additionally, we aimed to characterize the biological significance of these findings through pathway analysis and functional enrichment, as well as the construction of interaction networks between miRNAs and their target genes. This approach significantly contributes to understanding the underlying molecular mechanisms in MCI, as well as to identifying potential molecular markers for this condition.

In this work, we performed an integrative meta-analysis of miRNA expression profiles in plasma, serum, and extracellular vesicles, specifically exosomes, from subjects with MCI and control subjects. We identified eight consistently upregulated miRNAs (hsa-miR-23a-3p, hsa-miR-5001-5p, hsa-miR-6087, hsa-miR-4488, hsa-miR-6724-5p, hsa-miR-149-3p, hsa-miR-150-3p, and hsa-miR-99b-5p) and four consistently downregulated miRNAs (hsa-miR-4431, hsa-miR-1185-2-3p, let-7f-5p, and hsa-miR-550a-5p) by means of VCR method. The interaction network analysis between miRNA-mRNA highlighted the possible key role of hsa-miR-149-3p and Let-7f-5p in MCI.

The hsa-miR-149-3p was found as a central node with a high degree of connectivity in the interaction network of upregulated miRNAs (Figure 4A), suggesting its possible influence in regulating multiple cellular processes through the modulation of a wide range of target genes. Specifically, hsa-miR-149-3p was found to be upregulated in serum samples, indicating its potential as a minimally invasive diagnostic marker (Chen et al., 2008). This miRNA has been implicated in various pathological conditions, including several types of cancer such as glioma (He et al., 2018; Lu et al., 2022; Wang et al., 2023) and central nervous system diseases like amyotrophic lateral sclerosis (ALS) (Katsu et al., 2019). Additionally, it is involved in several critical biological processes such as cell proliferation, apoptosis, and cell cycle regulation (Wang et al., 2023). However, its potential influence on cognitive decline has not yet been established (Xiamin et al., 2020), and the expression of hsa-miR-149-3p in serum samples in MCI has not been validated.

Interestingly, a study reported a positive correlation between serum levels of hsa-miR-149, the precursor of hsa-miR-149-3p, and MMSE scores in Alzheimer’s disease (AD) patients, suggesting that miR-149 might be involved in the progression of this disease (Du et al., 2021). Given that the MMSE is widely used for assessing MCI, it is crucial to investigate whether this association also manifests in specific pathological processes of MCI and determine which of the two strands of the hsa-miR-149 precursor (hsa-miR-149-3p or hsa-miR-149-5p) could be specially involved in these conditions. Contrary to our observations in this meta-analysis, in AD patients this miRNA was observed downregulated. However, in another work, a similar pattern of expression was observed in patients with HIV-associated neurocognitive disorder, where the levels of hsa-miR-149-3p was found increased by comparing this group with HIV patients without this disorder (Asahchop et al., 2016). These findings support the idea that this miRNA might reflect the underlying biological alterations in neurocognitive disorders.

From the interaction network analysis in our study, we also determined that the consistently upregulated miRNAs, including hsa-miR-149-3p, exert significant influence in regulating key genes. We observed that NACC1 and FOXK1 were targets of hsa-miR-149-3p and proved to be important nodes (hubs) within the network. Both NACC1 and FOXK1 have been reported to act as transcriptional regulators (Gao et al., 2020; Schoch et al., 2017). Specifically, NACC1 is expressed in the CNS (Mackler et al., 2000; Xie et al., 2024) and plays a crucial role in maintaining synaptic plasticity (Daniel et al., 2023; Shen et al., 2007). On the other hand, FOXK1, whose DNA methylation has been associated with AD, has been linked to oxidative stress (De Jager et al., 2014; Sanchez-Mut and Gräff, 2015). These associations suggest that both NACC1 and FOXK1 could play significant roles in various biological processes related to MCI. However, further research is needed to confirm their direct involvement and to understand in more detail the specific mechanisms through which these genes and their interactions with hsa-miR-149-3p could influence the progression and pathogenesis of MCI. Additionally, it remains to be determined whether the expression of these genes can be indicative of pathological states related to MCI.

Regarding the consistently downregulated miRNAs, the interaction network identified Let-7f-5p as a central node with a high degree of connectivity (Figure 4B), showing consistent negative regulation in plasma and plasma-derived exosomes from subjects with MCI. It is important to note that the expression of Let-7f-5p has been reported differentially across various sample types in subjects with AD. For instance, Hara et al. (2017) reported upregulation of Let-7f-5p in serum samples from AD patients, whereas Leidinger et al. (2013) reported reduced levels of Let-7f-5p in whole blood. However, a study examining the association between plasma miRNA levels and changes in cognition over time found a negative correlation between Let-7f-5p and MMSE scores, suggesting that Let-7f-5p expression is linked to cognitive decline (Comfort et al., 2022). This indicates potential variation in its expression depending on the disease stage, ranging from pre-symptomatic or MCI stages to dementia or AD development. Nevertheless, further research is needed to confirm this relationship.

The variability in Let-7f-5p expression across different sample types underscores the need to consolidate these findings to establish its utility as a potential marker for MCI. Consolidating data will enable its potential clinical use, offering a non-invasive tool for early detection of MCI and monitoring progression to dementia or AD.

The interaction network analysis in our study revealed that consistently downregulated miRNAs, including Let-7f-5p, exert significant influence in regulating key genes, prominently featuring PMAIP1. This gene, recognized as an essential mediator of p53-dependent apoptosis, plays a crucial role in neurodegenerative disorders, particularly in cellular stress response (Gil-Jaramillo et al., 2023; Roufayel et al., 2022). Furthermore, a recent omics data integration study identified PMAIP1 as a potential risk gene for AD (Wang et al., 2021). These observations and the increased risk of progression to AD due to MCI open new perspectives for evaluating the possible relationship between PMAIP1 and MCI, which could be crucial for better understanding the underlying mechanisms in the transition from MCI to AD.

Additionally, we identified hsa-miR-23a-3p as a central node in the interaction network between upregulated miRNAs and differentially expressed genes in the peripheral blood of patients with amnestic MCI, suggesting its potential as a key regulator in the pathogenesis of MCI. Previously, hsa-miR-23a-3p has been reported as downregulated in plasma extracellular vesicles and serum from patients with AD (Galimberti et al., 2014; Gámez-Valero et al., 2019). However, studies in brain tissue from AD patients have shown upregulation of this miRNA (Lau et al., 2013), highlighting a discrepancy in its expression depending on the type of sample and the biological context in which it is analyzed. According to our results, hsa-miR-23a-3p was consistently reported as upregulated in serum and plasma samples from MCI patients, suggesting that its regulation may depend on factors specific to the disease stage and the particular biological environments of each sample. This discrepancy in the regulation patterns of hsa-miR-23a-3p could indicate a complex biological function across the different stages of the disease. In this context, further studies are imperative to investigate the functional role of hsa-miR-23a-3p in greater depth, in order to elucidate its contribution to the underlying pathological mechanisms in both MCI and AD.

Our KEGG analysis revealed that HIF-1 signaling pathway, p53 signaling pathway, protein processing in the endoplasmic reticulum, and focal adhesion pathways were consistently regulated by both groups of miRNAs (Figure 3). Considering these pathways’ critical roles in neuronal function and early stages of neurodegeneration (Rønn et al., 2000; Sin and Nollen, 2015), these findings suggest that the regulation of these pathways may be linked to pathological mechanisms associated with MCI. Moreover, the involvement of miRNAs in both repression and activation of these pathways highlights the diversity of their regulatory function and suggests a complex interaction of potential molecular mechanisms underlying MCI.

Finally, we found that both groups of miRNAs were associated with biological processes that could influence the regulation of metabolic pathways essential for neuronal function. Their involvement in the modification of cellular proteins and the regulation of gene expression highlights their potential role in modulating key processes for neuronal plasticity and cognition, which are critical factors for maintaining cognitive abilities (Bhat et al., 2022).

Considering that the upregulated miRNAs showed a stronger association with the stress response and cell death, it is possible that they are involved in the adaptive response to neuronal damage, contributing to cellular defense mechanisms and participating in the regulation of neuronal homeostasis in response to stress that may be occurring during MCI (Butterfield, 2023). On the other hand, the consistent downregulation of miRNAs associated with Toll-like receptor signaling pathways suggests an alteration in immune responses and the regulation of brain inflammation. These processes have been identified as key factors in the pathogenesis of MCI, indicating that the regulation of these pathways could have important implications for modulating neuroinflammation (Liang and Wang, 2021; Su et al., 2016).

Together, these results suggest that both upregulated and downregulated miRNAs play complementary roles in the pathophysiology of MCI. Understanding how these miRNAs influence these biological processes could open new avenues for the development of therapeutic strategies aimed at restoring cellular balance and mitigating the effects of neuroinflammation in MCI.

Although the meta-analyses play a crucial role in detecting miRNAs with more reliable differential expression by integrating multiple studies, thereby increasing sample size and overcoming limitations inherent in individual studies (Wang et al., 2015):these analyses may face certain limitations, such as the scarcity of studies evaluating miRNA profiles in subjects with MCI. Additionally, the lack of availability of raw data poses another obstacle to research. For example, while we identified studies in PubMed analyzing miRNA profiles in MCI subjects, many authors do not share raw data in any database. Therefore, it is essential for all studies to provide raw data on expression profiles, which would facilitate access to and reuse of these public datasets for future analyses (Forero et al., 2019).

Other limitation related to our work was the heterogeneity between the included studies, which could affect the robustness of the results obtained using the RRA method. Differences in the platforms and protocols used, along with the scarcity of studies evaluating miRNAs in patients with MCI, complicated the interpretation of the findings. However, the complementary VCR approach provided additional insights, helping to mitigate some of these challenges. Since research on miRNAs in MCI is an emerging field, we believe that addressing these limitations in future studies is essential for identifying the profile and functional role of miRNAs in MCI.

In general, our findings are promising, as they open new perspectives for future research that could lead to the identification of specific molecular markers for MCI in easily accessible biological samples, providing valuable insights into the underlying molecular mechanisms of this condition. Validating the expression of hsa-miR-149-3p, Let-7f-5p, and hsa-miR-23a-3p in plasma and serum samples, as well as elucidating their interactions with target genes, are crucial steps to confirm their clinical relevance in the context of MCI. This knowledge could be key to developing new therapeutic strategies aimed at modulating the expression of specific miRNAs or the activity of their target genes, with the goal of preventing, delaying, or reversing cognitive decline in individuals at risk of developing AD (Abidin et al., 2023; Li et al., 2024).

The identification of specific miRNA profiles for MCI, along with a detailed understanding of the biological mechanisms involved, has crucial implications not only for the diagnosis and treatment of this condition but also for optimizing clinical management and designing more effective and personalized therapeutic strategies. This approach would represent a significant step toward precision medicine in the field of neurodegenerative diseases (Nuzziello et al., 2020). Furthermore, the detection of miRNAs in blood samples, supported by advanced technologies such as next-generation sequencing and PCR, would facilitate their integration into daily clinical practice, improving both diagnostic accuracy and therapeutic outcomes. However, for this approach to be clinically viable, it is essential to validate the findings in a broader biological context, ensuring their applicability across diverse populations and clinical settings, thus consolidating their inclusion in routine clinical practice.

In summary, our results pave the way for promising future research that could bring these advances to fruition, transforming the diagnosis and management of MCI and other neurodegenerative diseases, and providing a solid foundation for future clinical applications.