Alzheimer’s disease (AD) is a common form of dementia that causes progressive mental decline in the elderly. This is due to impaired neurons and synapses, eventually changing patients’ behavior, memory, language, and cognition (Lu et al., 2021). AD is characterized by the accumulation of disposition of amyloid-β (Aβ), neurofibrillary tangle (NFT) formation, and extensive neuronal degeneration (Maciotta et al., 2013). The etiology of AD remains incompletely comprehended. The early indications of AD are frequently characterized by mild symptoms, including memory lapses and challenges in recollecting recent occurrences. As the disease advances, affected individuals may manifest symptoms such as cognitive impairment, perceptual disorientation, linguistic deficits, affective instability, and behavioral alterations.

The utilization of microRNA (miRNA) panels, which consist of multiple miRNAs, has been studied by researchers to enhance diagnostic accuracy. The panels typically consist of a group of differentially expressed miRNAs that, when combined, provide a more comprehensive disease profile. miRNA is a short non-coding RNA that plays an essential role in biological and pathological mechanisms in humans, animals, and plants (Harper et al., 2019; Toivonen et al., 2020; Beylerli et al., 2022). In recent years, many miRNAs have been studied and found to be involved with the development of neurodegenerative diseases, including AD. This makes them a potential biomarker for neurodegenerative diseases.

On the other hand, circular RNAs (CircRNAs) function as regulators of miRNA activity by sequestering them as miRNA sponges or competitive endogenous RNAs (ceRNAs) due to the presence of miRNA binding sites within circRNAs (Hansen et al., 2013a). When miRNAs form complexes with circRNAs, they undergo a process of sequestration or “sponging” from their respective target messenger RNAs (mRNAs) (Hansen et al., 2013b). Furthermore, circRNA regulates host gene expression through various mechanisms, such as acting as protein sponges, facilitating protein translocation and translation, and promoting protein-protein interaction (Zhao et al., 2022). This interaction can be especially important when the targeted miRNAs regulate biological processes or disease pathways. This review provides an overview of miRNAs and circRNAs’ biosynthetic mechanisms and functional roles. Additionally, the most relevant miRNAs and circRNAs, as well as the relationship miRNA-circRNA linked to AD, are discussed. Furthermore, our attention will be directed toward how the levels of multiple miRNAs in peripheral blood, serum, exosomes derived from serum, cerebrospinal fluid, and brain are altered in individuals with AD.

Most miRNAs are known to be synthesized via a canonical pathway, which undergoes several processing steps before developing into functional miRNAs. In the canonical pathway, primary miRNAs (pri-miRNAs) are transcribed by RNA polymerase II within the nucleus, which generates distinguish hairpin structures with polyadenylated and capped. The hairpin structures within pri-miRNAs are recognized by DGCR8 (DiGeorge syndrome critical region gene 8) and cleaved DROSHA to form precursor miRNAs (pre-miRNAs). The pre-miRNAs are exported to the cytoplasm by exportin 5 (XPO5) via Ran-GTP-dependent mechanisms (Komatsu et al., 2023).

The pre-miRNA is further cleaved off by dicer and its catalytic partner trans-activator RNA-binding protein (TRBP) into ∼22 nt miRNA duplexes. The miRNA duplex is loaded into Argonaute (AGO) protein before unwinding into two strands, guide, and passenger strand. Generally, the RNA-induced silencing complex’s component 3 promoter (C3PO) (RISC) will degrade the passenger strand. At the same time, the guide strand will be incorporated with AGO2 to initiate the RNA silencing mechanism. In some cases, both strands can be incorporated into the RISC complex. The RISC complex then recognizes target mRNAs based on sequence complementary between the miRNA and the mRNA 3′untranslated region (UTR). This recognition leads to mRNA degradation or translation inhibition, subsequently reducing protein expression.

CircRNAs exist as a continuous loop structure. circRNAs are ubiquitously expressed in mammalian cells and can also be found in a tissue-specific manner (Han et al., 2018). In contrast to linear mRNA, circRNA is generated through back-splicing, RNA-binding protein (RBP)-mediated circularisation, and exon skipping driven by lariat or intron-pairing mechanisms from pre-messenger RNA (Sekar and Liang, 2019; Guria et al., 2020). This process results in a circular transcript structure that arises from the fusion of the 3′non-co-linearly splice site with either its upstream 5′or another upstream exon during splicing events (Zhou et al., 2020).

The circRNA molecules that contain exonic or exon-intronic sequences are formed through a process known as back-splicing (Patop et al., 2019). CircRNAs can also be produced using lariat-driven circularisation forming a circular exonic RNA with a large lariat of introns (Guria et al., 2020). In addition, the RBP can bind specific motifs in the flanking introns to promote circularisation (Patop et al., 2019). Several splicing factors, including QK1, ADAR (adenine deaminase acting on RNA), ESRP1, and FUS, have been shown to facilitate circRNA production (Conn et al., 2015; Ivanov et al., 2015). Then circRNAs are transported into the cytoplasm in a size-dependent manner through either ATP-dependent RNA helicase DDX39A (also referred to as nuclear RNA helicase URH49) or spliceosome RNA helicase DDX39B (also known as DEAD box protein UAP56) (Huang et al., 2018).

Many studies have demonstrated that dysregulation of miRNAs was associated with the pathogenesis of AD. One of the studies identified a set of miRNAs differentially expressed in the brain, cerebrospinal fluid (CSF), and blood of AD patients compared to healthy controls (Takousis et al., 2019). Although miRNAs are predominantly intracellular, a sizeable fraction of them are migratory and can be identified in extracellular fluids (Zen and Zhang, 2012). These miRNAs are known as circulating miRNAs and can be detected in bodily fluids such as blood, urine, saliva, seminal fluid, breast milk, microvesicles, and exosomes (Weber et al., 2010; Vickers et al., 2011; O’Driscoll, 2015). This review presents a comprehensive summary of miRNA expression in three distinct compartments, blood, CSF, and brain tissues, in relation to AD (Table 1).

A new class of non-coding RNAs known as circRNAs has been identified. The circRNAs are produced by a non-canonical splicing event called back-splicing (Kristensen et al., 2019). This RNA was initially identified in 1976 as a viroid consisting of a single-stranded circular RNA molecule isolated from an infected tomato plant (Sanger et al., 1976). This RNA molecule exhibits a structure that is covalently closed-like structure, with a high degree of self-complementarity and base-pairing. The circRNAs have been identified as miRNA sponges due to having highly abundant miRNA binding sites (Hansen et al., 2013b). CircRNAs play a crucial role in developing and maintaining brain homeostasis in the brain. Several studies have shown that a healthy mammalian brain has the highest circRNA expression level and varies across different brain regions (Rybak-Wolf et al., 2015; You et al., 2015; Li et al., 2017).

The circRNAs were highly abundant in the cerebellum, followed by the prefrontal cortex and hippocampus (Rybak-Wolf et al., 2015). Interestingly, circRNAs expression was higher in neurons than in astrocytes in the cerebral cortex and cerebellum (Gokool et al., 2020). In neurons, they are highly concentrated in the synapses of the hippocampus, indicating their potential involvement in synaptic plasticity and cognitive processes (You et al., 2015). Many circRNAs have been shown to interact with disease-associated miRNAs, suggesting that circRNAs could play major roles in the development of diseases and as a prospective prognostic biological marker. For instance, a study by Zhao and co-workers showed that ubiquitin-protein ligase A (UBE2A) expression in AD is affected by the sponging activity of CDR1-as and miR-7 (Zhao et al., 2016). The miR-138-circHDAC9 complex is an additional mechanism that regulates the metabolism of Aβ in AD. The level of circHDAC9 was reduced in the serum of individuals with AD, which may lead to upregulation of miR-138 expression and downregulation of Sirt1 and ADAM10. Consequently, the processing of amyloid APP was redirected from the β-secretase pathway to the β-secretase pathway, resulting in an elevation of amyloid accumulation (Lu et al., 2019).

ciRS-7, the most characterized circRNA, comprises more than 70 conserved binding sites for miR-7 (Hansen et al., 2013b). The ciRS-7 is a robust and stable expression in various tissue, particularly in neural tissues (Ma et al., 2020). The downregulation of ciRS-7 has negatively correlated with miR-7 (Kristensen et al., 2019). The upregulation of miR-7 results in the downregulation of targets associated with AD, namely, UBE2A, which hinders the degradation of APP and β-secretase in the brain (Zhao et al., 2016). Frontotemporal lobar degeneration (FTLD-TAU) patients with 53 MAPT gene mutations are strongly associated with AD (Ghetti et al., 2015). Back-splicing exon 12 to 7 (tau circ 12→7) or 12 to 10 (tau circ 12 → 10), the MAPT gene produces two circRNAs in the brain, contributing to AD pathology. Only tau circ 12 → 7 has a start codon, but 12 → 10 circRNA may require ADAR activity to start translation (Welden et al., 2022). Despite the absence of stop codons in both circRNAs, protein translation can still occur at various locations in circular patterns, producing abnormal proteins. These circRNAs proteins tend to self-aggregate and form neurofibrillary tangles, possibly contributing to FTLD progression (Welden et al., 2022).

Another circRNA involved in tau phosphorylation is circPCCA. The circPCCA might competitively bind to miR-138-5p, inhibiting miR-138-5p′s ability to induce glycogen synthase kinase-3β activation and facilitate tau phosphorylation. These findings indicate that high circPCCA expression can potentially reduce disease severity in AD (Wang et al., 2015; Li et al., 2020).

Based on microarray analysis of CSF derived from AD patients, there was an upregulation of 112 circRNAs and a downregulation of 51 circRNAs when compared to the control subjects (Li et al., 2020). These circRNAs were enriched in pathways associated with AD, such as the neurotrophin signaling pathway, natural killer cell-mediated cytotoxicity, and cholinergic synapse. Further validation using RT-qPCR has demonstrated an increase in the expression of circLPAR1, circAXL, and circGPHN, while circPCCA, circHAUS4, circKIF18B, and circTTC39C were found to be decreased in AD patients.

Despite the intense research on miRNAs expression profiles for diagnosing AD, they are still in their early stage of development. One of the reasons is the complexity of the disorders to understand their pathological cause, which involves many different pathways and targets. Certain miRNAs displayed a distinct expression pattern specific to particular tissues or developmental stages. These miRNAs play a significant role in preserving tissue identity and function by contributing to various biological processes. Several miRNAs have been identified as brain-specific miRNAs, including miR-143, miR-125a/b, miR-138, miR-708, and miR-9 (Guo et al., 2014). Interestingly, these miRNAs have been identified in various other biological compartments, including blood, serum, plasma, and CSF (Table 1), which suggests that these miRNAs have the potential to be valuable candidates for biomarkers in AD in the future.

Moreover, based on the thorough assessment of the miRNAs expression profile (Figure 1), this present study identified several miRNAs also hold promise as a potential tool for AD diagnosis, monitoring disease progression, and gaining insights into the underlying molecular mechanisms of the disease. These miRNAs (Figure 1) were found to be associated with genes such as MAPK, BACE1, PTGS2, STAT3, SNAP25, and BDNF. This association subsequently contributes to the development of AD. For instance, the upregulation of miR-125 in AD promotes Tau hyperphosphorylation by regulating MAPK kinases (Banzhaf-Strathmann et al., 2014). This activation is most likely achieved by down-regulating the expression of phosphatase genes that miR-125 targets, which are DUSP6, Bcl-W, and PPP1CA (Banzhaf-Strathmann et al., 2014). BDNF, an additional gene implicated in AD, is linked to cognitive decline, particularly in immediate memory. The expression of miR-613 was found to have a negative regulatory effect on the expression of BDNF (Li et al., 2016). This relationship was observed in the hippocampus of APP/PS1 mice, where high levels of miR-613 coincided with low levels of BDNF expression (Li et al., 2016). In addition, the upregulation of miR-210-5p was observed to cause a reduction in the number of synapses in primary hippocampal neurons (Ren et al., 2018). Conversely, inhibition of miR-210-5p led to an increase in synaptic formation. This condition can be attributed to the downregulation of SNAP25, which is negatively inhibited by miR-210-5p (Ren et al., 2018).

There was a positive correlation between the expression of serum miR-331-3p and MMSE scores in patients with AD (Liu and Lei, 2021). The SH-SY5Y cell viability was significantly increased by overexpressing miR-331-3p while inhibiting miR-331-3p reduced cell viability. Meanwhile, the levels of proinflammatory cytokines, such as IL-1β, IL-6, and TNF-α, were decreased when miR-331-3p was overexpressed. Conversely, when miR-331-3p was knocked down, these cytokine levels increased. The treatment of Aβ resulted in a significant decrease in the expression of miR-331-3p. This finding aligns with the observed reduction of miR-331-3p in serum samples from individuals with AD (Liu and Lei, 2021).

STAT3 phosphorylation was markedly elevated in SH-SY5Y cells after transfection with an inhibitor of miR-29c-3p or miR-19b-3p (Wu et al., 2017). Based on computational predictions, miR-29c-3p, and miR-19b-3p target the same sequence in the 3′-UTR of STAT3. The hippocampus of the AD mice model and AD post-mortem brain shows substantial increases in phosphorylation of STAT3. In addition, STAT3 has served as a transcriptional regulator of BACE1, the crucial enzyme involved in producing Aβ (Millot et al., 2020). Other miRNAs that were observed to decrease their expression when the BACE1 level was increased were miR-16-5p and miR-19b-3p (Wu et al., 207). The overexpression of miR-16-5p or miR-19b-3p reduced the adverse effects of Aβ on cell viability and apoptosis in SH-SY5Y cells. Conversely, the knockdown of these miRNAs promoted the injury induced by Aβ.

In addition to miR-16-5p and miR-19b-3p, miR-107 and miR-384 also have complementary binding sites on the 3′UTR of BACE1, which leads to the suppression of its expression (Wang et al., 2008; Liu et al., 2014; Zhang et al., 2020). The relationship between miR-107 and circ_0049472 was confirmed through a dual-luciferase reporter and RNA pull-down assays. Circ_0049472 was found to be overexpressed in CSF and serum samples derived from AD patients. This gene was also overexpressed in SK-N-SH and CHP-212 cells induced with Aβ. The cell viability and apoptosis induced by the downregulation of circ_0049472 in Aβ-induced SK-N-SH and CHP-212 cells were abolished when miR-107 was silenced (Zheng et al., 2022).

Another gene that is related to AD is APP. A previous study discovered that miR-455 regulates APP by binding to its 3′UTR. This regulation was protective against mitochondrial and synaptic abnormalities induced by mutant APP in AD (Kumar et al., 2019). The PTGS2 was predicted to be targeted by miR-103, and it was discovered that PTGS2 is regulated in the opposite direction by miR-103 expression. In the PC12 cellular AD model, transfection with PTGS2 and miR-103 mimic plasmid decreased total neurite growth compared to the miR-103 mimic group. Additionally, there was an increase in cell apoptosis, suggesting that PTGS2 mimic mitigated the impact of miR-103 mimic on AD progression (Yang et al., 2018). In addition, circ_0000950 decreased the expression of miR-103 and increased the expression of PTGS2 in rat pheochromocytoma cell line PC12 cells and rat cerebral cortex neurons AD models. Overexpression of circ_0000950 resulted in increased neuron apoptosis, decreased neurite outgrowth, and elevated levels of IL-1β, IL-6, and TNF-α. On the other hand, the knockdown of circ_0000950 inhibited neuron apoptosis, promoted neurite outgrowth, and reduced levels of IL-1β, IL-6, and TNF-α. These effects were observed through the direct sponging of miR-103 (Yang et al., 2019).

Certain miRNAs, such as miR-455-3p and miR-210, have been found to possess protective mechanisms. According to Table 1, the expression of these miRNAs was upregulated and downregulated in AD patients, respectively. This condition has raised speculation that these miRNAs are protective in reducing AD severity. There are variations in the expression patterns of different miRNAs, such as miR-331-3p, miR-125-5p, and miR-384, across various biological sources. These changes in their environment potentially lead to their ability to respond, or they may also have a protective role in AD severity. Nevertheless, the precise mechanisms governing the expression of these microRNAs remain unidentified based on current knowledge.

circRNAs are known for their high conservation and structural stability. They have been found to play a crucial role in the onset and progression of numerous diseases. Thus, they could be valuable biomarkers or targets for therapy. Nevertheless, there are still barriers that must be addressed. This includes investigating how circRNAs interact with specific miRNAs, mRNAs, and proteins. By understanding these interactions, we can uncover a regulatory network that plays a role in AD development. To confirm the clinical relevance of circRNAs in AD, further investigation is required to understand their mechanisms and establish correlations. This will involve analyzing a large cohort of patient samples. Furthermore, there is an urgent need for technical advancements in accurately quantifying a specific circRNA and effectively silencing it without impacting the expression of the parent linear transcript.

Also, miRNAs can be detected in the bloodstream and potentially be used as biomarkers for the disease. It is essential to mention that miRNA can be detected in CSF. However, miRNA has the unique ability to traverse the blood-brain barrier and maintain its integrity, protected against degradation. This is achieved by its interaction with protein complexes and its containment within membrane-bound vesicles, such as exosomes (Hill, 2019). miRNA levels in circulation have the potential to indicate neuronal function and dysfunction accurately. This suggests they could be used as innovative therapeutic targets for treating dementia. Several recent meta-analyses have been conducted to establish a unified miRNA signature for AD. For example, in their study, Swarbrick et al. (2019) discovered a peripheral blood microRNA signature comprising ten molecules that could be linked to Braak Stage III.

In contrast, Bottero and Potashkin (2019) made predictions regarding miRNAs that were expected to influence the expression of genes known to exhibit differential expression in individuals with MCI and AD. Diagnostic tools and the significant heterogeneity among studies have limited the use of miRNAs as biomarkers and identifying specific miRNAs associated with AD. The heterogeneity observed in this context can be attributed to several factors, including variations in sample handling and the utilization of different profiling techniques such as microarrays, NGS, RT-qPCR, and other analytical approaches. Hence, further research is required to develop consistent protocols, determine dependable biomarkers, and comprehend the functional consequences of miRNA dysregulation in AD.

In summary, miRNA and circRNA represent promising avenues for developing non-invasive biomarkers for Alzheimer’s disease. The continued exploration of these non-coding RNA molecules in the context of AD has the potential to revolutionize AD diagnosis, monitoring, and therapeutic interventions, ultimately contributing to better disease management and improved quality of life for affected individuals.