Fecal microRNA patterns linked to Alzheimer's disease

Several hypotheses have been proposed to describe the pathology, including Aβ aggregation, tau protein hyperphosphorylation, mitochondrial dysfunction, and oxidative stress [26 –28]. In addition to these, novel theories such as prion-like propagation of protein aggregates [29], cerebral vasoconstriction and vascular contributions to cognitive decline [30], and the role of altered gamma oscillations in network dysfunction [31] have also been proposed. Emerging evidence also points to the hormone secretagogue receptor 1α (GHSR1α) and its influence on synaptic plasticity and neurodegeneration [32] as well as chronic infections such as Herpes simplex virus type 1, as potential contributors to the development of AD [33]. The deposition of Aβ is a hallmark feature of AD and plays a central role in the pathogenesis and progression of the disease. Aβ is derived from a larger protein called amyloid precursor protein (APP). In a healthy brain, APP undergoes processing by α-secretase or β-secretase, producing soluble, non-toxic fragments that are efficiently metabolised or cleared, preventing harmful accumulation. However, in AD, APP is abnormally cleaved by β-secretase (BACE-1), followed by γ-secretase, leading to the formation of Aβ peptides, primarily Aβ40 and Aβ42. Among these, Aβ42 is more prone to aggregation, resulting in the formation of neurotoxic oligomers, which subsequently cluster into amyloid plaques within the CNS. These plaques contribute to synaptic dysfunction, neuroinflammation, and neuronal death, ultimately driving the progression of AD. An elevated Aβ42/Aβ40 ratio, together with increased total Aβ levels, is widely recognised as a key contributor to AD pathogenesis and is strongly linked to a heightened risk of developing the condition [34].

NFTs are composed of hyperphosphorylated tau protein aggregated into paired helical filaments [35]. The accumulation of tau disrupts the microtubule structure essential for axonal transport, thereby impairing intracellular trafficking and mitochondrial function. This impacts synaptic integrity and neuronal viability, playing a significant role in the neurodegenerative cascade characteristic of AD [36].

Swerdlow, Burns, and Khan [37] suggest that mitochondrial dysfunction is a pivotal early event in the progression of AD. Impaired mitochondrial function results in reduced cellular energy production, which in turn initiates a cascade of neurodegenerative processes. Key consequences include the excessive generation of reactive oxygen species (ROS), disruption of calcium homeostasis, and dysfunction in mitochondrial quality control mechanisms such as mitophagy. Collectively, these alterations contribute to neuronal damage, synaptic loss, and ultimately, cognitive decline in AD [38].

Oxidative stress typically arises from the presence of free radicals and ROS, which are formed by molecules containing unpaired electrons from oxygen- and nitrogen-based compounds. In AD, the deposition of Aβ leads to the binding of these proteins to mitochondrial membranes, disrupting their function and prompting the production of ROS. This overabundance of ROS within neuron cell bodies exacerbates the condition, initiating processes such as elevated intracellular calcium levels, mitochondrial dysfunction, and DNA damage linked to oxidative stress [39].

Emerging evidence suggests that dysregulated fecal miRNAs are involved in key pathological processes in AD, including Aβ aggregation, tau hyperphosphorylation, neuroinflammation, and synaptic dysfunction (64–65). These miRNAs, derived from both host and gut microbiota, represent promising non-invasive biomarkers for early detection and monitoring of AD, with the added benefit of reflecting gut–brain axis disruptions. Recent investigations have revealed that specific fecal miRNAs are differentially expressed in AD, with distinct patterns of downregulation and upregulation (66–68). Downregulated miRNAs include miR-29a, which negatively regulates β-site amyloid precursor protein cleaving enzyme 1 (BACE1) involved in Aβ production, and miR-132, a key regulator of synaptic plasticity and neuronal survival (69). miR-128 has also been shown to be reduced in AD and may influence the abundance of Akkermansia muciniphila, a beneficial gut microbe, potentially exacerbating gut dysbiosis (70). In contrast, miR-106b, miR-146a, and miR-155 are found to be upregulated in AD and are involved in pro-inflammatory pathways and microglial activation (69). miR-223 has been linked to immune regulation through modulation of Clostridium species, while miR-9 contributes to epithelial integrity and has been associated with the regulation of Bacteroides fragilis, a bacterium implicated in AD-associated inflammation (71–73). These regulatory miRNAs may serve as key intermediaries in the gut–brain axis, impacting both microbial composition and host immune responses. miR-223, for example, enhances host defence by modulating neutrophil activation and reducing Bacteroides overgrowth, which has been linked to amyloid accumulation (71). Similarly, miR-128 reduces Enterobacteriaceae abundance and suppresses pro-inflammatory cytokines (70), while miR-9 helps maintain gut epithelial integrity, influencing the prevalence of SCFA-producing bacteria such as Faecalibacterium (73). Together, these miRNAs represent potential molecular links between microbial ecology and neurodegeneration. Further research is necessary to validate these candidate biomarkers across larger cohorts and to elucidate causal pathways through which fecal miRNAs modulate AD pathophysiology (74–75).

Alterations in CSF biomarkers, such as the Aβ42/Aβ40 ratio and phosphorylated tau (p-tau) levels, are closely linked to miRNA expression changes in AD. Reduced Aβ42/Aβ40 ratio correlates with increased levels of miR-125b and miR-146a, which drive neuroinflammation and impair Aβ clearance by downregulating neprilysin (NEP) and insulin-degrading enzyme (IDE) [76]. Similarly, elevated p-tau levels are associated with increased miR-132 downregulation, leading to disrupted synaptic plasticity and neuronal loss [77]. Inflammatory cytokines like IL-6 and TNF-α also influence miRNA expression, with high levels of these cytokines linked to upregulation of miR-155, further exacerbating neuroinflammation and tau pathology [78].

Emerging evidence supports a dynamic and bidirectional relationship between gut microbiota and host miRNA expression, with important implications for the pathogenesis and progression of AD. Gut microbes and their metabolites including SCFAs, secondary bile acids, and LPS can modulate host miRNA expression both locally in the gut epithelium and systemically within the CNS [79]. SCFAs such as butyrate, produced mainly by Faecalibacterium prausnitzii, Roseburia spp., and Eubacterium spp., upregulate neuroprotective miRNAs like miR-124 and miR-132, which support synaptic function and suppress neuroinflammation [79, 80]. Conversely, host-derived miRNAs can directly regulate microbial composition; miR-146a, miR-515-5p, and miR-1226-5p secreted into the intestinal lumen via exosomes can enter bacterial cells and alter gene expression, with miR-1226-5p promoting Escherichia coli growth and miR-515-5p suppressing Fusobacterium nucleatum [81].

Several miRNAs appear to mediate gut–brain interactions relevant to AD. miR-223 and miR-146a regulate intestinal barrier integrity and are induced by dysbiosis [82], while miR-155 and miR-21, often elevated in response to endotoxins from Bacteroides fragilis or Proteobacteria such as Escherichia/Shigella, are increased in both the gut and the AD brain [83, 84]. These miRNAs promote NF-κB activation, microglial reactivity, and cytokine release, contributing to early inflammatory stages of AD. SCFA-producing bacteria can also reduce the expression of pro-inflammatory miRNAs, whereas the overgrowth of pathobionts such as Bilophila wadsworthia and Desulfovibrio spp. may elevate miRNAs associated with barrier dysfunction and inflammation [85]. However, some findings challenge existing assumptions; for example, elevated butyrate levels have been reported in certain AD models despite concurrent microglial activation and cognitive decline, suggesting SCFAs may not always exert protective effects [86].

Despite these advances, important gaps and inconsistencies remain. miR-132 shows a robust decline in the AD brain, yet peripheral levels often remain unchanged, raising concerns about whether circulating miRNAs reliably reflect CNS pathology [87]. Similarly, Akkermansia muciniphila is linked to improved cognition in some cohorts but increased inflammation in others, pointing toward strain-specific or host-dependent effects [88]. Temporal mapping is also lacking; it remains unknown when microbiota–miRNA alterations arise across the AD continuum, as most studies are cross-sectional and cannot establish causality. Tissue-specific divergence presents another challenge, as it is unclear how miRNA signatures differ between gut, blood, and brain in response to microbial shifts, and which compartment best reflects AD-related pathophysiology. Additionally, sex-based differences in microbial composition and miRNA regulation are underexplored, despite their potential to explain cohort variability [65].

Reduced abundances of gut microbes such as Rikenellaceae, unidentified Ruminococcaceae, and Alistipes, along with decreased microbial diversity, have been associated with higher risks of mild cognitive impairment (MCI) in AD [89]. These microbial shifts parallel reductions in key serum miRNAs, including hsa-let-7g-5p, hsa-miR-107, and hsa-miR-186-3p, which collectively suggest an early systemic link between gut dysbiosis and cognitive decline [86, 111]. Several fecal and circulating miRNAs show strong relevance to AD pathology. Reduced levels of miR-132, an essential regulator of synaptic plasticity and neuronal survival, have been reported in AD and may impair synaptic integrity, memory consolidation, and dendritic growth through dysregulation of CREB signalling [90, 91]. Experimental restoration of miR-132 has been shown to reduce tau hyperphosphorylation and neuronal degeneration, highlighting its therapeutic potential [92, 93]. miR-29a, which suppresses BACE1 and limits Aβ formation, is also downregulated in AD, supporting its role in amyloid accumulation and neurotoxicity [3, 4]. Meanwhile, dysregulation of other miRNAs, including miR-9, crucial for neurogenesis and neuronal connectivity, has been linked to disrupted synaptic function and cognitive processing [94, 95].

In contrast, several pro-inflammatory miRNAs are elevated in AD. miR-155 is consistently upregulated in fecal and peripheral samples and is strongly associated with microglial activation, cytokine release, neuroinflammation, and synaptic injury [93 –97]. miR-146a and miR-106b, also increased in AD, contribute to inflammatory signalling and impaired Aβ clearance, partly by suppressing neprilysin and insulin-degrading enzyme [96, 98]. Regulatory miRNAs such as miR-223 and miR-128 additionally shape gut microbiota composition and gut barrier integrity by modulating species like Clostridium, Bacteroides fragilis, Enterobacteriaceae, and SCFA-producing bacteria, including Faecalibacterium [7 –9]. Microbial metabolites—including short-chain fatty acids (SCFAs), secondary bile acids, and LPS—further influence host miRNA expression. SCFAs such as butyrate, produced by Faecalibacterium prausnitzii, Roseburia, and Eubacterium, enhance neuroprotective miRNAs like miR-124 and miR-132, whereas host miRNAs such as miR-1226-5p and miR-515-5p can directly regulate bacterial gene expression in species like E. coli and Fusobacterium nucleatum [4, 11]. Despite these insights, inconsistencies remain regarding SCFA levels, peripheral miRNA reliability, and species-specific microbial effects—underscoring the need for longitudinal, multi-tissue, and sex-stratified studies to clarify temporal dynamics and improve the diagnostic and therapeutic potential of fecal miRNAs in AD [5, 99 –102].

A growing body of evidence supports the mechanistic role of miRNAs in driving core pathological processes in AD, reinforcing their biological plausibility. One of the most consistently altered miRNAs, miR-132, regulates synaptic plasticity, dendritic growth, and neuronal survival, and its loss disrupts the CREB signalling pathway, which is central to memory consolidation and long-term potentiation [3, 90, 91, 103]. Downregulation of miR-132 leads to impaired CREB-dependent transcription and reduced expression of synaptic proteins necessary for neuronal connectivity. Experimental restoration of miR-132 in AD models rescues synaptic function and reverses dendritic atrophy, demonstrating a direct causal link between miR-132 deficiency and synaptic degeneration [104 –107].

Beyond synaptic dysfunction, miR-132 also influences tau phosphorylation and clearance, providing mechanistic insight into neurofibrillary tangle formation. Reduced miR-132 enhances the activity of tau kinases such as GSK-3β and CDK5, promoting tau hyperphosphorylation and aggregation [93, 94]. miR-132 deficiency further suppresses pathways responsible for tau degradation, including autophagy and proteasomal processing, thereby accelerating intracellular tau accumulation [8]. Studies show that reintroduction of miR-132 reduces phosphorylated tau and improves neuronal survival, emphasising its role in tau homeostasis [4, 5, 10, 11].

Multiple miRNAs, particularly miR-29a/b and miR-107, exert direct regulatory control over BACE1, the β-secretase responsible for amyloidogenic APP cleavage [3, 98]. Downregulation of miR-29 family members increases BACE1 expression, leading to excessive Aβ42 generation and early amyloid plaque deposition [107]. Similarly, reduced miR-107 expression correlates with increased BACE1 activity and accelerated disease progression, highlighting the role of miRNA-mediated β-secretase regulation in Aβ pathology [76, 77, 107]. Loss of miR-132 also indirectly enhances amyloidogenic processing, reinforcing its role as a central upstream regulator [104 –107].

Pro-inflammatory miRNAs such as miR-155 and miR-146a further contribute to AD pathology through modulation of neuroimmune signaling. Upregulation of miR-155 amplifies NF-κB–mediated cytokine production by suppressing SOCS1, promoting a sustained inflammatory state in microglia [72 –74, 117]. miR-146a, conversely, acts as a feedback regulator but becomes dysregulated in chronic inflammation, leading to abnormal suppression of IRAK1 and TRAF6 and altering innate immune responses in AD [78]. The presence of these inflammatory miRNAs in both gut and brain tissues links gut dysbiosis to central neuroinflammatory processes, demonstrating a mechanistic interface between peripheral and CNS pathology [79, 80].

Diet plays a central role in shaping gut microbial communities, and these microbial shifts directly influence host miRNA expression relevant to AD pathology [79]. Fibre-rich and anti-inflammatory diets promote the expansion of beneficial SCFA-producing taxa such as Faecalibacterium prausnitzii, Roseburia spp., and Eubacterium spp., which subsequently enhance the expression of neuroprotective miRNAs including miR-124 and miR-132 [79, 80]. Conversely, Western-style diets high in saturated fats and refined sugars promote dysbiosis characterised by increases in Escherichia/Shigella, Bacteroides fragilis, Bilophila wadsworthia, and Desulfovibrio spp., leading to elevated LPS and other inflammatory metabolites that upregulate miR-155, miR-21, and miR-146a [81 –85]. These observations demonstrate that dietary patterns shape microbial ecology in ways that directly modulate the host miRNA environment [76 –79].

Diet-induced microbial alterations also influence key miRNAs implicated in AD-related molecular pathways, including amyloidogenesis, tau phosphorylation, and synaptic plasticity [90, 91, 103]. SCFA-associated microbial profiles support the maintenance of miR-132 and miR-29a/b, which regulate BACE1 expression and CREB-mediated neuronal survival, thereby protecting against amyloid accumulation and synaptic loss [3, 98]. Dysbiosis caused by unhealthy dietary patterns suppresses these miRNAs, leading to increased BACE1 activity, enhanced Aβ production, and dysregulated tau kinase activity that accelerates neurofibrillary tangle development [104 –107]. In parallel, inflammatory diets elevate miR-155 and miR-21, enhancing NF-κB activation and microglial reactivity, which amplify neuroinflammation and exacerbate cognitive decline [82 –84, 97]. Thus, diet-driven microbial dysbiosis modulates miRNA networks that directly feed into established AD molecular mechanisms [90].

Despite increasing evidence supporting the role of gut microbiota and fecal miRNA in AD, several controversies and unresolved questions remain. A major debate concerns the dualistic effects of butyrate, a key SCFA. While butyrate is widely recognised for its neuroprotective actions, including anti-inflammatory effects, modulation of microglial activation, and upregulation of beneficial miRNAs such as miR-124 and miR-132 [59 –61]. Other studies have reported paradoxical findings. In some AD models, SCFA levels, including butyrate, are elevated despite worsening pathology, suggesting that the impact of butyrate may be highly context dependent [86, 99]. These discrepancies indicate that butyrate's effects may vary based on microbial composition, host metabolism, disease stage, and receptor sensitivity, highlighting the need for more mechanistic research to clarify whether SCFAs are uniformly protective or capable of contributing to pathology under specific conditions.

Another prominent controversy involves the strain-specific and host-dependent roles of Akkermansia muciniphila in neurodegeneration. Although A. muciniphila has been widely reported to enhance mucosal integrity, reduce inflammation, and support SCFA metabolism,mechanisms thought to be beneficial in AD [52 –55].Other studies indicate contradictory outcomes. Some reports show that particular strains of A. muciniphila may exacerbate inflammation or metabolic stress depending on host factors [88, 100]. This suggests that its role in AD is complex, potentially beneficial in some physiological or dietary contexts while detrimental in others. These inconsistencies emphasise the importance of analysing A. muciniphila at the strain-level, rather than generalising its effects across all variants, and point toward a need for host–microbe interaction studies that incorporate immune status, diet, and microbial ecology.

A final unresolved issue concerns the discrepancies between fecal and brain miRNA profiles in AD. While fecal miRNAs provide a valuable window into gut–brain axis dysregulation, inflammatory states, and microbial shifts, their expression patterns do not always mirror those found in the CNS. For instance, miR-132 and miR-29a are consistently downregulated in AD brain tissue and are closely linked to synaptic failure, tau hyperphosphorylation, and impaired Aβ regulation [90 –92, 98]. However, several studies report inconsistent, minimal, or unchanged levels of these same miRNAs in fecal or circulating samples [87, 99]. These discrepancies likely arise from tissue-specific regulation, differential release mechanisms, microbial interactions, and degradation differences between brain and stool environments. Additionally, technical factors such as RNA extraction variability, library preparation inconsistencies, and bioinformatic pipeline differences may further amplify these conflicting results [78, 79]. Overall, resolving these inconsistencies will require multi-tissue, longitudinal research directly comparing miRNA signatures from gut, circulation, and brain across preclinical, MCI, and dementia stages.

Personalised miRNA-based therapies tailored to individual gut microbiome profiles represent an emerging direction in precision medicine for AD, as both microbial composition and host miRNA signatures show strong inter-individual variability linked to cognitive outcomes [89, 90]. Integrating microbiome sequencing with miRNA profiling allows identification of patient-specific dysregulated pathways, such as reduced miR-132, miR-29a, and miR-107 or elevated miR-155, miR-146a, and miR-21, which directly influence amyloid processing, tau phosphorylation, neuroinflammation, and synaptic plasticity [3, 87, 98,]. This approach also enables targeted restoration of protective miRNAs that decline in AD such as miR-132, miR-124, and miR-128 or suppression of pathogenic ones, including miR-155 and miR-21, according to each patient's molecular profile [91 –93, 103]. Because gut microbial communities strongly shape host miRNA expression, incorporating individualised microbiome data may guide interventions aimed at modifying dysbiosis-associated miRNAs, especially those responsive to microbial metabolites such as SCFAs and LPS [79 –82]. Precision strategies can further reduce off-target effects by aligning therapeutics with each patient's unique miRNA–microbiota network, thereby improving safety and enhancing long-term disease modification potential [104 –108].

Although microRNA-based therapies show strong preclinical promise in AD, their translation to humans remains limited by the difficulty of achieving efficient and targeted delivery across the BBB, even when using engineered miRNA mimics or inhibitors [104 –107]. Off-target effects represent another major challenge because individual miRNAs regulate multiple gene networks simultaneously, increasing the likelihood of unintended alterations in synaptic, inflammatory, or metabolic signalling pathways [3, 98, 109]. Translational uncertainty is further increased by substantial differences between rodent models and human neurobiology, raising concerns about whether therapeutic strategies—such as restoring miR-132 can reliably replicate their beneficial effects when applied clinically in AD patients [104 –107]. Additional complexity arises from the dynamic behaviour of disease-associated miRNAs, including miR-155, miR-146a, and miR-21, which shift in response to gut dysbiosis and systemic inflammation, making long-term therapeutic modulation in humans unpredictable [82 –84, 97].