How do pro-inflammatory cytokines alter miRNA profiles in brain endothelial cells? Mechanisms, consequences, and therapeutic strategies for blood-brain barrier protection

1 Introduction

The blood-brain barrier (BBB) is a highly specialized neurovascular structure that regulates molecular and cellular exchange between the bloodstream and the central nervous system (CNS). Brain microvascular endothelial cells form the anatomical basis of the BBB, characterized by tight junctions, adherens junctions, and specialized transport systems that restrict paracellular permeability (Abbott et al., 2010). These intercellular junctions, composed of transmembrane proteins including claudins, occludin, and junctional adhesion molecules, are anchored to the cytoskeleton through scaffolding proteins such as zonula occludens (ZO) proteins (Tietz and Engelhardt, 2015). During infectious diseases and neuroinflammatory conditions, the BBB becomes compromised, facilitating pathogen entry, immune cell infiltration, and disruption of CNS homeostasis (Daneman and Prat, 2015; Varatharaj and Galea, 2017; Sweeney et al., 2019) (Figure 1).

Pro-inflammatory cytokines, particularly TNF-α and IFN-γ, are key mediators of BBB dysfunction during infection and inflammation (Chaitanya et al., 2011; O'Carroll et al., 2015; Rochfort and Cummins, 2015). These cytokines are produced by activated immune cells and can directly act on brain endothelial cells, altering their phenotype and barrier properties (Huynh and Dorovini-Zis, 1993; Maeda et al., 2021). TNF-α, a pleiotropic cytokine, increases paracellular permeability in a dose-dependent manner and enhances pathogen penetration across the BBB (Fiala et al., 1997). IFN-γ induces profound morphological and functional changes in brain microvessel endothelial cells, including the surface expression of class II major histocompatibility complex antigens and increased permeability to macromolecules (Chaitanya et al., 2011).

Recent evidence has revealed that microRNAs (miRNAs)—small, non-coding RNAs of approximately 22 nucleotides that post-transcriptionally regulate gene expression-play a critical role in mediating the effects of cytokines on brain endothelial cells (Reijerkerk et al., 2013; Cerutti et al., 2016; Wang et al., 2021; Sun et al., 2022). miRNAs bind to complementary sequences in the 3’ untranslated regions of target mRNAs, leading to translational repression or mRNA degradation (Bartel, 2009). In the context of the BBB, miRNAs have emerged as versatile regulators of endothelial barrier function under both physiological and pathological conditions (Lopez-Ramirez et al., 2014; Yin et al., 2014; Wu and Chen, 2016; Cerutti et al., 2017). Altered levels of BBB-associated miRNAs have been demonstrated in numerous neurological diseases characterized by loss of inter-endothelial junctions and disruption of barrier integrity (Wu and Chen, 2016).

This review describes how TNF-α and IFN-γ modulate miRNA expression profiles in human brain endothelial cells and how these changes contribute to the molecular events that lead to BBB disruption during infection. We synthesize evidence from in vitro models using human cerebral microvascular endothelial cell lines (hCMEC/D3) and primary human brain microvascular endothelial cells, as well as in vivo studies in animal models of neuroinflammation and infection (Lopez-Ramirez et al., 2013; Weksler et al., 2013). Understanding these regulatory mechanisms is essential for developing targeted therapeutic strategies to preserve BBB integrity and prevent neurological complications in infectious and inflammatory CNS diseases.

2 Molecular mechanisms of BBB disruption by TNF-α and IFN-γ

2.1 Cytokine-signaling pathway activation in brain endothelial cells

TNF-α and IFN-γ activate multiple intracellular signaling pathways in brain endothelial cells, thereby mediating their effects on barrier function and inflammatory responses. The NF-κB pathway is a central mediator of cytokine-induced endothelial activation (Pahl, 1999). TNF-α binding to its receptors triggers the activation of the IκB kinase (IKK) complex, leading to the phosphorylation and degradation of IκB proteins and the subsequent nuclear translocation of NF-κB (Hayden and Ghosh, 2008). Once in the nucleus, NF-κB induces the transcription of pro-inflammatory genes, including adhesion molecules, chemokines, and cytokines (Hayden and Ghosh, 2008).

Recent studies have demonstrated that inflammatory reprogramming of brain endothelial cells through NF-κB activation compromises BBB integrity (Pahl, 1999). An NF-κB-associated endothelial gene module is elevated in neuroinflammatory conditions and enriched for inflammation and endothelial-to-mesenchymal transition signatures. Cytokine stimulation of brain endothelial cells induces morphological remodeling, lipid accumulation, and junctional disruption, all of which can be prevented by NF-κB inhibition. This highlights NF-κB as a critical node in cytokine-induced BBB dysfunction and a potential therapeutic target (Pahl, 1999).

Mitogen-activated protein kinase (MAPK) pathways, including extracellular signal-regulated kinase (ERK), JNK, and p38 MAPK, are also activated by TNF-α and IFN-γ in brain endothelial cells (Hayden and Ghosh, 2008). These pathways regulate diverse cellular processes, including gene transcription, protein synthesis, cell survival, and apoptosis (Pinals et al., 2025). In the context of BBB dysfunction, MAPK activation contributes to tight junction disruption, adhesion molecule expression, and matrix metalloproteinase (MMP) production (Pinals et al., 2025). For example, TNF-α-induced MMP-9 expression in human cerebral microvascular endothelial cells is mediated by the Ca²⁺/CaMKII/ERK/NF-κB signaling pathway (Kyriakis and Avruch, 2012).

Glycogen synthase kinase-3β (GSK3β) has emerged as an important regulator of cytokine-induced BBB dysfunction. GSK3β inhibition restores barrier integrity and reduces adhesion molecule expression in TNF-α-stimulated brain endothelial cells (Pinals et al., 2025). The protective effects of GSK3β inhibition are mediated in part through modulation of miRNA expression and activation of the Wnt/β-catenin signaling pathway. miR-27a-3p, for instance, induces expression of claudin-5 and occludin by downregulating GSK3β and activating the Wnt/β-catenin signaling pathway. This pathway is essential for BBB maintenance, and its disruption contributes to barrier dysfunction in neurological disorders (Rosenberg, 2009; Cargnello and Roux, 2011; Ding et al., 2019; Harati et al., 2022).

Tissue-nonspecific alkaline phosphatase (TNAP) activity in brain endothelial cells promotes the maintenance of barrier integrity via the ROCK pathway. Treatment with TNF-α and IFN-γ significantly decreases endothelial TNAP activity and barrier integrity. Loss of TNAP activity contributes to BBB dysfunction through cytoskeletal remodeling and increased ROCK 1/2 expression. This represents another mechanism by which pro-inflammatory cytokines compromise BBB integrity through modulation of endothelial cell signaling pathways (Nwafor et al., 2021).

2.2 Tight junction protein modulation

TNF-α and IFN-γ disrupt BBB integrity through multiple mechanisms that converge on tight junction protein complexes. These cytokines negatively modulate cell-cell contacts by altering the expression and localization of key junctional molecules, including claudin-5, occludin, and ZO-1 (Nitta et al., 2003; Maeda et al., 2021). Claudin-5, the most abundant tight junction protein in brain endothelial cells, is particularly susceptible to cytokine-mediated downregulation. Studies using in vitro BBB models have demonstrated that TNF-α treatment causes a concentration-dependent decrease in claudin-5 expression and altered localization, with recovery time dependent on cytokine concentration (Lopez-Ramirez et al., 2012).

The mechanisms underlying cytokine-induced disruption of tight junctions involve both transcriptional and post-transcriptional regulation. At the transcriptional level, TNF-α and IFN-γ activate signaling pathways that suppress the expression of tight junction protein genes. At the post-transcriptional level, cytokine-induced miRNAs directly target tight junction protein mRNAs, leading to their degradation or translational repression (Engelhardt and Ransohoff, 2012). For example, miR-155 targets claudin-1, a tight junction protein that contributes to barrier integrity, resulting in its downregulation and subsequent reorganization of ZO-1 at cell-to-cell junctions (Nitta et al., 2003; Wong et al., 2007; Lopez-Ramirez et al., 2012; Ge et al., 2019; Maeda et al., 2021).

In addition to reducing tight junction protein expression, cytokines induce the redistribution of junctional proteins from cell-cell contacts to the cytoplasm. This redistribution is mediated by the activation of signaling pathways, including c-Jun N-terminal kinase (JNK) and protein kinase C (PKC), which phosphorylate tight junction proteins and promote their internalization. Studies have shown that combined TNF-α and IFN-γ treatment causes the redistribution of ZO-1 and VE-cadherin, which disrupts both tight junctions and adherens junctions. These changes result in increased paracellular flux and compromised barrier function (Persidsky et al., 1999; Engelhardt and Ransohoff, 2012; Ge et al., 2019).

The role of caspases in cytokine-induced barrier breakdown has also been elucidated. Lower concentrations of TNF-α and IFN-γ activate caspase-3 and caspase-9, causing increased paracellular permeability and junctional complex redistribution without inducing apoptosis (Persidsky et al., 1999). Higher cytokine concentrations fully activate caspase-3/7, leading to endothelial cell apoptosis and complete barrier breakdown. This dose-dependent effect highlights the importance of cytokine levels in determining the extent and reversibility of BBB dysfunction (Persidsky et al., 1999; Engelhardt and Ransohoff, 2012).

2.3 Adhesion molecule expression and leukocyte trafficking

TNF-α and IFN-γ promote leukocyte adhesion and transmigration across the BBB by upregulating endothelial adhesion molecules. These cytokines significantly increase the expression of ICAM-1, vascular cell adhesion molecule-1 (VCAM-1), and E-selectin on brain endothelial cells (Suárez et al., 2010; Yun et al., 2019). The upregulation of adhesion molecules facilitates the multi-step process of leukocyte recruitment, including rolling, firm adhesion, and transendothelial migration (Suárez et al., 2010; Yun et al., 2019).

The process of leukocyte transmigration itself contributes to BBB disruption. As leukocytes migrate across the endothelium, they disrupt junctional complexes and increase paracellular permeability (Engelhardt and Ransohoff, 2012). TNF-α enhances this process by increasing both leukocyte adhesion and endothelial barrier permeability, thereby creating a paracellular route for immune cell infiltration (Engelhardt and Ransohoff, 2012). Studies using in vitro BBB models have demonstrated that TNF-α stimulates brain endothelial cells to show increased monocyte adhesion and migration, which is associated with the upregulation of adhesion molecules and inflammatory chemokines (Suárez et al., 2010; Yun et al., 2019).

Brain endothelial cells also release microparticles in response to inflammatory cytokine stimulation, which may contribute to barrier dysfunction and leukocyte trafficking. These microparticles carry proteins from BBB junctions, including claudin-1, claudin-3, claudin-5, occludin, and VE-cadherin. Cytokine-induced apical microparticles (AMPs) impair BBB endothelial solute permeability, while basolateral microparticles (BMPs) affect smooth muscle contractility. The release of these distinct microparticle populations is regulated by separate molecular pathways involving Rho-associated protein kinase (ROCK), calpain, and cholesterol depletion (Rochfort et al., 2014; O'Carroll et al., 2015; Wu and Chen, 2016; Stephan et al., 2022).

3 Cytokine-induced miRNA expression changes in brain endothelial cells

3.1 TNF-α-induced miRNA alterations

TNF-α exposure induces substantial changes in the miRNA landscape of brain endothelial cells. Comprehensive miRNA profiling studies have revealed that TNF-α stimulation of human brain microvascular endothelial cells results in the downregulation of 123 miRNAs and the upregulation of 193 miRNAs. Among the most consistently reported TNF-α-induced miRNAs are members of the let-7 family and miR-98, which are upregulated in primary human brain microvascular endothelial cells following TNF-α exposure. These miRNAs target inflammatory molecules, including CCL2 (MCP-1) and CCL5 (RANTES), suggesting a negative feedback mechanism to limit excessive inflammation (Lopez-Ramirez, 2012; Lopez-Ramirez et al., 2013).

In peripheral endothelial cells, TNF-α induces specific miRNAs that regulate adhesion molecule expression. Studies in human umbilical vein endothelial cells (HUVECs) have shown that TNF-α upregulates miR-31 and miR-17-3p, which target E-selectin and intercellular adhesion molecule-1 (ICAM-1), respectively. While these studies were not conducted in brain endothelial cells, they provide important mechanistic insights into TNF-α-mediated miRNA regulation of endothelial activation that may be conserved across vascular beds. Additionally, miR-125a-5p has been shown to be upregulated in brain endothelial cells, where it downregulates TNF-α-mediated ICAM-1 expression and reduces monocyte transmigration (Lopez-Ramirez, 2012; Lopez-Ramirez et al., 2013).

3.2 IFN-γ and combined cytokine effects

IFN-γ, either alone or in combination with TNF-α, produces distinct miRNA expression profiles in brain endothelial cells. The combination of TNF-α and IFN-γ is particularly relevant to neuroinflammatory conditions, as these cytokines are often co-expressed during immune responses in the CNS. Combined cytokine treatment (TNF-α + IFN-γ) induces a more pronounced inflammatory phenotype in brain endothelial cells compared to either cytokine alone, with corresponding changes in miRNA expression (Lopez-Ramirez, 2012; Lopez-Ramirez et al., 2013).

One of the most extensively studied cytokine-responsive miRNAs in brain endothelial cells is miR-155, which is strongly and rapidly upregulated by the combination of TNF-α and IFN-γ. miR-155 plays a central role in modulating brain endothelial barrier function during neuroinflammation by targeting multiple components of cell-cell junctional complexes and focal adhesion structures. Transcriptomic analysis has revealed that miR-155 overexpression in human cerebral microvascular endothelial cells downregulates 390 transcripts, with significant enrichment in biological processes related to translation, focal adhesion, and inter-endothelial junctional complexes (Lopez-Ramirez et al., 2014; Cerutti et al., 2016; Barker et al., 2017).

Another important cytokine-responsive miRNA is miR-146a, which is upregulated in brain endothelial cells following treatment with TNF-α and IFN-γ. miR-146a expression is elevated in microvessels of multiple sclerosis (MS) active lesions and in the spinal cord of experimental autoimmune encephalomyelitis (EAE) mice (Wu et al., 2015). Unlike miR-155, which promotes barrier dysfunction, miR-146a appears to serve a protective role by negatively modulating T-cell adhesion to brain endothelial cells through the repression of multiple targets, including RhoA, nuclear factor of activated T cells 5 (NFAT5), IL-1 receptor-associated kinase 1 (IRAK1), and TNF receptor-associated factor 6 (TRAF6). This leads to the inhibition of NF-κB activation and decreased expression of adhesion molecules (Wu et al., 2015).

The differential effects of TNF-α alone versus combined TNF-α and IFN-γ treatment highlight the complexity of cytokine-miRNA networks in brain endothelial cells. Cytokine-induced changes in gene expression profiles of hCMEC/D3 cells reveal that combined treatment produces synergistic effects on barrier permeability, cell-to-matrix contacts, and transporter systems. These changes are partly mediated by alterations in miRNA expression, which in turn modulate the ability of cerebral endothelial cells to respond to inflammatory stimuli (Lopez-Ramirez et al., 2013).

3.3 Temporal dynamics of miRNA responses

The temporal dynamics of cytokine-induced miRNA changes in brain endothelial cells are critical for understanding the progression from acute inflammatory responses to chronic barrier dysfunction. Studies examining time-course responses have identified distinct phases of miRNA regulation following cytokine exposure. Early-phase responses (within 6–12 hours) are characterized by the rapid upregulation of inflammation-responsive miRNAs such as miR-155 and miR-146a, which serve as immediate modulators of inflammatory signaling pathways (Lopez-Ramirez et al., 2014; Wu et al., 2015; Barker et al., 2017).

Intermediate-phase responses (12–48 hours) involve the sustained expression of early-response miRNAs along with the emergence of additional miRNAs that regulate junctional proteins and barrier integrity. During this phase, the balance between barrier-disrupting and barrier-protective miRNAs determines the extent of BBB dysfunction. Late-phase responses (beyond 48 hours) may involve compensatory mechanisms, including the upregulation of barrier-protective miRNAs such as let-7 and miR-98, which aim to restore barrier function and limit excessive inflammation (Lopez-Ramirez et al., 2014; Rom et al., 2015; Bernstein et al., 2020). The temporal pattern of miRNA responses is influenced by several factors, including cytokine concentration, duration of exposure, and the presence of other inflammatory mediators. High concentrations of cytokines or prolonged exposure may overwhelm compensatory mechanisms, leading to sustained barrier dysfunction and potential endothelial cell apoptosis. Understanding these temporal dynamics is essential for identifying optimal therapeutic windows for miRNA-based interventions (Lopez-Ramirez et al., 2014; Wu et al., 2015; Barker et al., 2017).

4 Integrative framework: cytokine-specific, context-specific, and disease-specific miRNA effects

We propose a three-level framework that distinguishes (i) cytokine-specific, (ii) context-specific, and (iii) disease-specific effects. The same miRNA may belong to more than one level, and the level at which it has been characterized dictates how its findings should be interpreted.

4.1 Cytokine-specific effects

This level captures miRNAs whose expression in brain endothelial cells is directly modulated by TNF-α and/or IFN-γ in defined in vitro systems, independent of the disease context. miR-155, miR-146a, the let-7/miR-98 family, miR-125a-5p, and miR-let-7i are paradigmatic examples: each is induced or repressed within hours of cytokine exposure in hCMEC/D3 cells or primary human brain microvascular endothelial cells; has a characterized signaling input (predominantly NF-κB for miR-155/miR-146a, and STAT1-co-operative signaling for miR-98/let-7); and has a validated target in junctional or adhesion-molecule mRNAs (Lopez-Ramirez et al., 2013; Lopez-Ramirez et al., 2014; Rom et al., 2015; Wu et al., 2015; Barker et al., 2017; Bernstein et al., 2020). For these miRNAs, the cytokine–miRNA link is mechanistically anchored in brain endothelial cells.

4.2 Context-specific effects

The direction and magnitude of cytokine-induced miRNA changes depend on cytokine concentration, duration of exposure, co-stimulation, and the vascular bed examined. Low-dose TNF-α primarily induces reversible tight-junction reorganization in which miR-155 plays a leading role, whereas high-dose or prolonged TNF-α/IFN-γ co-exposure shifts the response toward caspase-mediated barrier breakdown and a partially distinct miRNA signature (Persidsky et al., 1999; Engelhardt and Ransohoff, 2012; Lopez-Ramirez et al., 2012). Several adhesion-regulating miRNAs (miR-31, miR-17-3p) were identified in HUVECs and other peripheral endothelial cells and have not yet been formally validated in brain endothelium, illustrating that the vascular bed itself is a context. Temporal dynamics – early induction of miR-155/miR-146a versus delayed induction of let-7/miR-98 (Lopez-Ramirez et al., 2014; Rom et al., 2015; Wu et al., 2015; Bernstein et al., 2020) – are also context-specific and shape whether the net cytokine effect on the barrier is disruptive or protective. Findings at this level should not be treated as universal cytokine effects on brain endothelium.

4.3 Disease-specific effects

In pathological states, TNF-α/IFN-γ signaling is combined with additional pathogen- or tissue-derived signals that reshape the endothelial miRNA landscape. Cerebral malaria adds sequestration-driven endothelial activation and parasite-derived extracellular vesicles to the cytokine background (Mantel et al., 2016; Dunst et al., 2017; Zuniga et al., 2022); ischemic stroke superimposes hypoxia, which itself induces miR-210 and miR-212/132 (Ma et al., 2017; Yan et al., 2022); multiple sclerosis combines chronic IFN-γ exposure with autoimmune T-cell signaling (Wu et al., 2015). The same miRNA can therefore appear barrier-protective in one disease and barrier-disrupting in another. In subsequent sections we apply this framework: Section 5 primarily evaluates barrier-disrupting and barrier-protective miRNAs at the cytokine- and context-specific levels, while Section 6 addresses the disease-specific level in cerebral malaria.

5 miRNA-mediated regulation of BBB integrity

Many of the miRNAs reviewed in this section were originally identified in models of ischemic stroke, multiple sclerosis, traumatic brain injury, or hypoxia rather than in pure TNF-α/IFN-γ stimulation of brain endothelial cells. We include them here because each of these conditions involves substantial TNF-α and IFN-γ signaling as part of their endothelial pathophysiology, and the majority of the miRNAs discussed have either been directly shown to respond to these cytokines in brain or peripheral endothelial cells or to act on cytokine-regulated junctional and adhesion-molecule targets. For each miRNA we indicate whether the link to TNF-α/IFN-γ has been established directly in brain endothelial cells, extrapolated from cytokine-driven studies in other vascular beds, or remains correlative. The framework introduced in Section 4 (cytokine-specific, context-specific, and disease-specific levels) is applied throughout.

5.1 Barrier-disrupting miRNAs

Several miRNAs have been identified as negative regulators of BBB integrity, with their upregulation contributing to barrier dysfunction during neuroinflammation. miR-155 is the most extensively characterized barrier-disrupting miRNA in brain endothelial cells. As discussed earlier, miR-155 is strongly upregulated by TNF-α and IFN-γ and targets multiple components of cell-cell junctional complexes and focal adhesion structures. Specific validated targets of miR-155 include annexin-2 (ANXA2), claudin-1 (CLDN1), dedicator of cytokinesis 1 (DOCK1), and syntenin-1 (SDCBP). Overexpression of miR-155 mimics cytokine-induced alterations in junctional organization and permeability, while loss of miR-155 in knockout mice reduces BBB leakage in models of neuroinflammation (Lopez-Ramirez et al., 2014; Wu et al., 2015; Barker et al., 2017; Wei et al., 2022).

miR-210 has been identified as another barrier-disrupting miRNA in the context of neonatal hypoxic-ischemic brain injury (Ma et al., 2017), a setting in which hypoxia and TNF-α/NF-κB signaling converge, so that its relevance to cytokine-driven barrier injury is inferred by extrapolation rather than demonstrated directly in TNF-α/IFN-γ-stimulated brain endothelium. miR-210 suppresses junction proteins and disrupts BBB integrity by targeting multiple genes involved in tight junction assembly and maintenance (Ma et al., 2017). Similarly, miR-143 has been shown to compromise BBB integrity in the context of methamphetamine abuse. Silencing miR-143 preserves BBB integrity, suggesting that its upregulation contributes to barrier dysfunction (Bai et al., 2016).

The miR-15a/16–1 cluster has been implicated in BBB dysfunction during ischemic stroke. Endothelium-targeted deletion of the miR-15a/16–1 cluster ameliorates BBB dysfunction, indicating that these miRNAs negatively regulate barrier integrity (Ma et al., 2020). The mechanisms involve targeting of genes that maintain endothelial junctions and regulate vascular permeability (Ma et al., 2020). Additionally, miR-494-3p, particularly when modified by N6-methyladenosine (m6A), compromises the BBB by targeting tight junction protein 1 (TJP1) in intracranial atherosclerosis (Woudenberg et al., 2024). It should be noted that miR-143 (methamphetamine exposure), the miR-15a/16–1 cluster (ischemic stroke), and miR-494-3p (intracranial atherosclerosis) were characterized in disease models that are not primarily cytokine-driven; they are included here because they act on the same junctional targets engaged by TNF-α/IFN-γ, but a direct link to these cytokines in brain endothelium remains to be established. They therefore illustrate context-specific rather than cytokine-specific effects in the framework of Section 4.

Hypoxia-induced miR-212/132 alters BBB integrity through inhibition of tight junction-associated proteins in human and mouse brain microvascular endothelial cells. These miRNAs are upregulated under hypoxic conditions and contribute to barrier dysfunction by targeting genes involved in junction assembly and stability. The identification of these barrier-disrupting miRNAs provides potential therapeutic targets for preventing or reversing BBB dysfunction in neurological diseases (Yan et al., 2022).

5.2 Barrier-protective miRNAs

In contrast to barrier-disrupting miRNAs, several miRNAs have been identified as positive regulators of BBB integrity. These barrier-protective miRNAs are often downregulated during neuroinflammation, and their restoration can ameliorate barrier dysfunction. miR-98 and let-7g are among the most well-characterized barrier-protective miRNAs. As mentioned earlier, these miRNAs are induced in primary human brain microvascular endothelial cells after TNF-α exposure and target inflammatory molecules CCL2 and CCL5. Overexpression of let-7 and miR-98 reduces leukocyte adhesion and migration, diminishes pro-inflammatory cytokine expression, and increases BBB tightness, attenuating barrier leakiness under neuroinflammatory conditions (Rom et al., 2015; Bernstein et al., 2020).

miR-98 has been further validated as a barrier-protective miRNA in mouse models of ischemia/reperfusion stroke. miR-98 reduces endothelial dysfunction by protecting BBB integrity and improving neurological outcomes. Protective effects are mediated through targeting of genes involved in inflammatory responses and endothelial activation (Rom et al., 2015; Bernstein et al., 2020).

miR-126-3p and miR-126-5p are endothelial-enriched miRNAs that play important roles in maintaining vascular integrity. Overexpression of miR-126-3p/-5p attenuates BBB disruption in a mouse model of middle cerebral artery occlusion. These miRNAs regulate endothelial cell proliferation, migration, and angiogenesis, and their protective effects on the BBB involve maintaining tight junction protein expression and reducing inflammatory responses (Cerutti et al., 2017; Chen et al., 2026).

miR-29a-5p alleviates traumatic brain injury-induced disruption of the blood-brain barrier via regulation of the NLRP3 pathway. miR-29a-5p targets NLRP3 through binding to its 3’ UTR and negatively regulates its expression. Overexpression of miR-29a-5p decreases endothelial cell permeability, increases transendothelial electrical resistance (TEER), and enhances the protein content of ZO-1 and occludin (Pan et al., 2020). miR-29b similarly suppresses inflammation and protects BBB integrity in ischemic stroke (Zhang et al., 2021).

Additional barrier-protective miRNAs include miR-27a-3p, which regulates the expression of intercellular junction proteins at the brain endothelium and controls endothelial barrier permeability (Harati et al., 2022). miR-130a regulates cerebral ischemia-induced BBB permeability by targeting Homeobox A5 (Li, 2022), while miR-150 regulates BBB permeability via Tie-2 after permanent middle cerebral artery occlusion (Wang et al., 2018). miR-199a-5p attenuates BBB disruption following ischemic stroke by regulating PI3K/Akt signaling (Fang et al., 2016), and miR-132/212 regulates tight junction stabilization in the BBB after stroke (Ni et al., 2024). Of these protective miRNAs, only miR-98/let-7 and miR-146a have been shown to respond directly to TNF-α and/or IFN-γ in brain endothelial cells; miR-126, miR-29a-5p, miR-130a, miR-150, and miR-199a-5p were characterized chiefly in stroke and traumatic-injury models, and their responsiveness to these cytokines in brain endothelium is inferred rather than demonstrated. We retain them because they converge on the junctional and inflammatory pathways through which TNF-α and IFN-γ act, but this distinction should be kept in mind.

5.3 Feedback regulatory mechanisms

The regulation of BBB integrity by miRNAs involves complex feedback mechanisms that attempt to maintain homeostasis during inflammatory challenges. One important feedback mechanism involves the induction of barrier-protective miRNAs in response to inflammatory cytokines. As discussed, TNF-α induces both barrier-disrupting miRNAs (such as miR-155) and barrier-protective miRNAs (such as let-7 and miR-98). This suggests that endothelial cells mount compensatory responses to limit excessive inflammation and preserve barrier function (Rom et al., 2015; Bernstein et al., 2020).

The temporal dynamics of these feedback mechanisms are critical. Early upregulation of barrier-disrupting miRNAs may be necessary to initiate immune responses and enable controlled leukocyte trafficking into the CNS. However, sustained expression of these miRNAs can lead to pathological barrier dysfunction. The delayed upregulation of barrier-protective miRNAs may represent an attempt to restore homeostasis and prevent excessive barrier breakdown (Chen et al., 2026).

miRNAs also participate in feedback regulation of inflammatory signaling pathways. For example, miR-146a is induced by NF-κB activation and subsequently targets IRAK1 and TRAF6, key components of the NF-κB signaling pathway (Wu et al., 2015). This creates a negative feedback loop that limits excessive NF-κB activation and inflammatory responses. Similarly, TNF-α induces miRNAs that target adhesion molecules, providing feedback control of leukocyte recruitment (Suárez et al., 2010).

The balance between barrier-disrupting and barrier-protective miRNAs determines the net effect on BBB integrity. In healthy conditions or during mild inflammation, barrier-protective miRNAs may predominate, maintaining barrier function despite inflammatory challenges. However, in severe or chronic inflammation, barrier-disrupting miRNAs may overwhelm protective mechanisms, leading to sustained barrier dysfunction (Barker et al., 2017). Understanding these feedback regulatory mechanisms is essential for developing therapeutic strategies that tip the balance toward barrier protection (Ni et al., 2024).

Extracellular vesicles, including exosomes, serve as important mediators of intercellular communication and can transfer miRNAs between cells. Bone marrow stromal cell-derived exosomes containing miR-21a-5p ameliorate BBB injury and hemorrhagic transformation in ischemic stroke. These exosomes upregulate the expression of claudin-5, occludin, and ZO-1, and their protective effects are mediated through miR-21a-5p targeting of inflammatory pathways. This highlights the potential for exosome-based delivery of therapeutic miRNAs to protect BBB integrity (Yang et al., 2026).

6 BBB disruption in cerebral malaria: implications for miRNA-based therapy

Cerebral malaria (CM), caused by Plasmodium falciparum infection, represents a severe neurological complication characterized by profound BBB dysfunction. The pathogenesis of CM involves sequestration of infected erythrocytes in brain microvasculature, endothelial activation, and intense inflammatory responses. Cytokines, particularly TNF-α and IFN-γ, play central roles in CM pathogenesis, and their effects on brain endothelial cells are partly mediated by miRNA dysregulation (Storm and Craig, 2014; Dunst et al., 2017).

Studies comparing the effects of P. falciparum-infected red blood cells and TNF-α on human brain endothelial cells have revealed differential regulation of inflammatory and barrier-integrity pathways. While TNF-α induces high levels of endothelial activation with only a small increase in permeability, P. falciparum-infected red blood cells cause a marked increase in brain endothelial cell permeability, preferentially inducing an endoplasmic reticulum stress response. This suggests that parasite-induced signaling, distinct from TNF-α effects, is a major driver of BBB disruption in CM (Zuniga et al., 2022; Metwally et al., 2024).

miR-155 has been specifically implicated in CM pathogenesis. Upon stimulation with TNF-α and IFN-γ, miR-155 is upregulated in human brain endothelial cells, leading to reorganization of tight junction protein ZO-1 and increased endothelial permeability (Lopez-Ramirez et al., 2014; Barker et al., 2017). Claudin-1 is a candidate target of miR-155 in this context (Barker et al., 2017). Additionally, P. falciparum-infected erythrocyte-derived extracellular vesicles contain miR-451a, which correlates with reduced endothelial caveolin-1 expression upon endocytosis by endothelial cells (Mantel et al., 2016).

6.1 Sequestration-driven versus cytokine-only endothelial activation

Cytokine-only stimulation of brain endothelial cells in vitro produces a relatively uniform pro-inflammatory phenotype: NF-κB and STAT1 activation, upregulation of ICAM-1/VCAM-1, modest tight junction reorganization, and the canonical miR-155/miR-146a response described in Section 3 (Pahl, 1999; Lopez-Ramirez et al., 2013; Lopez-Ramirez et al., 2014; Barker et al., 2017). Sequestration of Plasmodium falciparum-infected red blood cells (iRBCs) in the cerebral microvasculature in cerebral malaria adds at least three distinct inputs that cytokine-only models do not capture. First, cytoadherence is mediated by direct binding of PfEMP1 variants to endothelial receptors such as EPCR and ICAM-1, which by itself activates intracellular signaling, suppresses the cytoprotective protein C pathway, and primes the endothelium independently of TNF-α (Dunst et al., 2017; Zuniga et al., 2022). Second, iRBCs trigger a marked endoplasmic reticulum stress response and an increase in paracellular permeability that exceeds what is observed with TNF-α alone, even though TNF-α produces a comparable or stronger activation transcriptome (Zuniga et al., 2022). Third, the spatial confinement of cytokine release at sites of sequestration creates locally extreme cytokine concentrations that are not reproduced in homogeneous in vitro cytokine exposure (Metwally et al., 2024). Functionally, this means that the miR-155-driven tight-junction remodeling characterized in TNF-α/IFN-γ-stimulated hCMEC/D3 cells (Barker et al., 2017) represents only one component of the cerebral malaria endothelial response; parasite-specific signaling contributes additional miRNA changes that are not phenocopied by cytokine stimulation alone; this needs more investigation from researchers.

6.2 Joint regulation of endothelial miRNAs by parasite-derived extracellular vesicles and host cytokines

Parasite-derived extracellular vesicles (EVs) and host cytokines act as parallel and partially overlapping inputs to brain endothelial miRNA networks. iRBC-derived EVs are taken up by endothelial cells in vitro and transfer miR-451a and other Ago2–miRNA complexes, with downstream effects including reduced caveolin-1 expression and altered barrier function (Mantel et al., 2016). These EV-delivered miRNAs do not require cytokine signaling to enter the endothelial cell, but their downstream targets overlap with cytokine-responsive pathways: caveolin-1 modulates eNOS and ICAM-1 signaling, both of which are TNF-α-regulated. EVs from infected erythrocytes and from cerebral malaria patients are also enriched in miR-let-7i, miR-150, and miR-146a (Chamnanchanunt et al., 2017; Joshi et al., 2024), which are independently induced by TNF-α/IFN-γ in brain endothelial cells (Lopez-Ramirez et al., 2014; Wu et al., 2015), suggesting that the same miRNAs can be delivered exogenously by parasites and induced endogenously by host cytokines. Comparative transcriptomic profiling of brain and lung endothelial cells exposed to P. falciparum-infected red blood cells further indicates that EV-delivered miRNAs and cytokine-induced miRNAs produce distinct but partially overlapping mRNA targetomes, with the combined response producing barrier disruption that neither stimulus reproduces alone (Metwally et al., 2024). The combinatorial nature of this regulation has practical therapeutic implications: neutralization of iRBC-EVs would not necessarily reverse the host-cytokine-driven miRNA component, and vice versa.

6.3 Concordance between in vitro BBB miRNA signatures and human cerebral malaria

A central translational question is whether the miRNA changes identified in cytokine-stimulated in vitro BBB models are reproduced in human cerebral malaria samples. The evidence is mixed. miR-155, the most robustly cytokine-induced miRNA in hCMEC/D3 cells (Lopez-Ramirez et al., 2014; Barker et al., 2017), is consistently elevated in experimental cerebral malaria (ECM) in mice: circulating exosomal miR-155 rises over the course of ECM, miR-155 deficiency preserves BBB integrity, and antagomir-155 reduces vascular leak in microvessels exposed to sera from Ugandan children with cerebral malaria (Barker et al., 2017; Martin-Alonso et al., 2018). In cross-sectional human cohorts, however, plasma or EV miR-155 has not consistently distinguished cerebral malaria from uncomplicated malaria, with several studies reporting non-significant or directionally inconsistent results (Chamnanchanunt et al., 2017; Joshi et al., 2024). miR-451a is reliably enriched in iRBC-derived EVs and correlates with severity in several cohorts (Mantel et al., 2016; Joshi et al., 2024; Gupta et al., 2025), but is abundant in erythrocytes generally, which complicates its specificity as a cerebral malaria marker. miR-21-5p, miR-126, and miR-150 each have plausible mechanistic links to cytokine signaling and barrier function, and elevated plasma miR-21-5p and miR-3158-3p correlate with cerebral malaria severity in an Indian adult cohort (Gupta et al., 2025; Gupta and Gupta, 2026); their behavior in pediatric African cohorts is less well established. Overall, miRNA signatures from in vitro cytokine-stimulated BBB models are partially – but not fully – reproduced in human cerebral malaria samples, with strong cross-system concordance for miR-155 in mechanistic murine studies but weaker concordance for circulating miR-155 as a clinical biomarker.

6.4 Inconsistencies between experimental models, human cohorts, and populations

Several factors contribute to the discrepancies between miRNA signatures identified in experimental cerebral malaria and those measured in human clinical samples, and these discrepancies have direct consequences for which candidates merit inclusion in Table 1 (Gupta et al., 2021a). First, experimental cerebral malaria in CBA or C57BL/6 mice infected with P. berghei ANKA is driven by CD8+ T-cell-dependent, IFN-γ-rich pathology that does not fully recapitulate human cerebral malaria, in which sequestration of iRBCs and platelet/monocyte accumulation are dominant features. miR-155 is illustrative: it is a robust mediator of BBB dysfunction in murine ECM (Barker et al., 2017), but findings in human samples are less consistent, and a detailed analysis of pediatric cerebral malaria did not identify miR-155 among the top discriminating miRNAs (Gupta et al., 2021b). Second, human cohorts differ substantially in age (pediatric African versus adult Asian), parasite burden, comorbidities (HIV, anemia, sickle hemoglobin), and sampling timing relative to disease onset, all of which influence circulating miRNA profiles; miR-451a is again illustrative, varying with hemoglobin genotype as well as with disease severity (Gupta et al., 2025). Similar inconsistencies apply to miR-146a and miR-27a, where SNP frequencies in pediatric cerebral malaria did not distinguish severe from uncomplicated disease (Wah et al., 2019). Third, the compartment of measurement matters: tissue or microvascular miRNA changes may not be reflected in plasma, total EVs, or peripheral blood mononuclear cells, all of which have been used as proxies in clinical studies. These considerations argue against direct extrapolation of murine ECM miRNA candidates to human cerebral malaria and motivate the redesigned Table 1, in which each miRNA is explicitly annotated with its experimental system, validation type, and clinical relevance.

Table 1. Standardized summary of miRNA candidates in cerebral malaria and Plasmodium infections, with explicit annotation of the experimental system, target gene, BBB relevance, validation type, clinical relevance, expression direction, and technique(s) used.

References are the same as in the main text.

Table 1 summarizes the key miRNA candidates implicated in cerebral malaria pathogenesis, including their expression patterns, experimental validation status, and molecular targets. Notably, miR-155 emerges as a central mediator of BBB disruption across multiple models (Barker et al., 2017; Martin-Alonso et al., 2018), whereas miR-126 downregulation correlates with disease severity (El-Assaad et al., 2011). These findings underscore the potential for miRNA-based biomarkers and therapeutic interventions in cerebral malaria.

Identifying the miRNAs responsible for BBB disruption has significant implications for miRNA-based therapy. First, recognizing the specific miRNAs that mediate cytokine-induced BBB dysfunction provides potential therapeutic targets. Inhibition of miR-155 or other barrier-disrupting miRNAs could preserve BBB integrity during CM and other neuroinflammatory conditions. Second, the discovery that parasites transfer miRNAs to endothelial cells via extracellular vesicles suggests that blocking this transfer or neutralizing parasite-derived miRNAs could be effective therapeutic strategies. Third, the differential effects of parasite-induced signaling versus cytokine signaling highlight the need for combination therapies that address multiple pathogenic mechanisms.

Therapeutic delivery of barrier-protective miRNAs represents another promising approach. As demonstrated in other contexts, exosomal delivery of miRNAs, such as miR-21a-5p can ameliorate BBB injury. Similar strategies could be applied to CM, where delivery of barrier-protective miRNAs (such as miR-98, let-7, or miR-146a) might counteract the effects of inflammatory cytokines and preserve BBB integrity. The timing of such interventions is critical, as early treatment before severe barrier breakdown occurs is likely to be most effective (Figure 2).

7 Discussion and future directions

The evidence synthesized in this review demonstrates that TNF-α and IFN-γ induce specific miRNA changes in human brain endothelial cells that play critical roles in mediating BBB dysfunction during infection and neuroinflammation. These cytokine-induced miRNA alterations affect multiple aspects of endothelial biology, including tight junction protein expression, adhesion molecule upregulation, activation of inflammatory signaling pathways, and leukocyte trafficking. The identification of both barrier-disrupting miRNAs (such as miR-155, miR-210, miR-143, and miR-15a/16-1) and barrier-protective miRNAs (such as miR-98, let-7, miR-146a, miR-126, miR-29a/b, and miR-27a-3p) provides a framework for understanding the molecular mechanisms of BBB dysfunction and for developing targeted therapeutic interventions (Rom et al., 2015; Bai et al., 2016; Fang et al., 2016; Ma et al., 2017; Wang et al., 2018; Bernstein et al., 2020; Ma et al., 2020; Pan et al., 2020; Zhang et al., 2021; Li, 2022; Yan et al., 2022; Ni et al., 2024; Woudenberg et al., 2024; Chen et al., 2026) (Figure 2).

Several key themes emerge from the current literature. First, the temporal sequence of cytokine-induced miRNA responses (Section 3.3) and the feedback networks that counter-regulate them (Section 5.3) together determine whether the net effect on the barrier is disruptive or protective and define the therapeutic window for intervention.

Second, the effects of cytokines on brain endothelial cells are context-dependent and are influenced by factors such as cytokine concentration, duration of exposure, presence of other inflammatory mediators, and the specific disease context. For example, the differential effects of P. falciparum-infected red blood cells versus TNF-α on brain endothelial cells highlight the importance of pathogen-specific mechanisms in addition to cytokine-mediated effects (Oxendine Harp et al., 2022). Fourth, miRNAs can be transferred between cells via extracellular vesicles, providing a mechanism for the remote modulation of brain endothelial cell function (Wah et al., 2019). This opens new avenues for both understanding disease pathogenesis and developing therapeutic delivery strategies.

The therapeutic potential of targeting miRNAs to preserve or restore BBB integrity is substantial. Several approaches can be envisioned: (1) Inhibition of barrier-disrupting miRNAs using antisense oligonucleotides (antagomiRs) or locked nucleic acid (LNA) inhibitors. miR-155 represents a particularly attractive target given its central role in mediating cytokine-induced barrier dysfunction. (2) Delivery of barrier-protective miRNAs using synthetic mimics or exosome-based delivery systems. The success of exosomal miR-21a-5p in ameliorating BBB injury in stroke models provides proof of concept for this approach. (3) Modulation of upstream signaling pathways that regulate miRNA expression, such as NF-κB or GSK3β. Inhibition of these pathways can shift the balance toward barrier-protective miRNA expression (Lopez-Ramirez, 2012; Rom et al., 2015; Wu et al., 2015; Bai et al., 2016; Fang et al., 2016; Wu and Chen, 2016; Barker et al., 2017; Ma et al., 2017; Wang et al., 2018; Ge et al., 2019; Bernstein et al., 2020; Ma et al., 2020; Pan et al., 2020; Zhang et al., 2021; Li, 2022; Wei et al., 2022; Yan et al., 2022; Ni et al., 2024; Woudenberg et al., 2024; Chen et al., 2026).

Several challenges must be addressed to translate miRNA-based therapies from preclinical studies to clinical applications. First, delivery of miRNA therapeutics to brain endothelial cells in vivo remains challenging due to the BBB, which restricts the entry of many therapeutic agents into the CNS. Strategies to overcome this barrier include using nanoparticles, exosomes, or focused ultrasound to enhance delivery (Bereczki et al., 2025). Second, off-target effects of miRNA therapeutics must be carefully evaluated, as miRNAs can regulate hundreds of target genes. Cell-type-specific delivery approaches may help minimize off-target effects (82). Third, the optimal timing and duration of miRNA-based interventions need to be determined for different disease contexts.

Future research directions should include: (1) Comprehensive characterization of miRNA expression profiles in brain endothelial cells across different disease contexts, including viral, bacterial, and parasitic infections, as well as sterile neuroinflammation. Single-cell RNA sequencing approaches could provide insights into the heterogeneity of miRNA responses within endothelial cell populations. (2) Mechanistic studies to identify and validate additional miRNA targets relevant to BBB function. While several targets have been identified (claudin-1, DOCK-1, SDCBP, ANXA-2, RhoA, NFAT5, IRAK1, TRAF6, NLRP3, GSK3β, and TJP1), many more likely exist. (3) Investigation of interactions between miRNAs and other non-coding RNAs, including long non-coding RNAs (lncRNAs) and circular RNAs (circRNAs), which may function as competing endogenous RNAs (ceRNAs) to modulate miRNA activity (82).

Development and testing of miRNA-based therapeutics in preclinical models of CNS infections and neuroinflammatory diseases. This should include an evaluation of different delivery methods, dosing regimens, and combination therapies. (5) Identification of miRNA biomarkers in blood or cerebrospinal fluid that could serve as indicators of BBB dysfunction and guide therapeutic interventions. Circulating miRNAs have shown promise as biomarkers in various neurological diseases (Gupta et al., 2021a). (6) Investigation of species differences in miRNA regulation of BBB function, as the majority of studies have used mouse models or human cell lines. Comparative studies could help identify conserved mechanisms that are most likely to translate to human disease.

The integration of miRNA biology with other emerging areas of BBB research, such as the roles of pericytes and astrocytes in barrier maintenance, the contribution of metabolic dysfunction to barrier breakdown, and the impact of the gut-brain axis on neuroinflammation, will provide a more comprehensive understanding of BBB regulation in health and disease (Gupta et al., 2021b). Systems biology approaches that model the complex interactions between cytokines, miRNAs, signaling pathways, and cellular responses will be essential for identifying key regulatory nodes that can be targeted therapeutically.

8 Outstanding questions

Despite significant advances in understanding cytokine-induced miRNA dysregulation in brain endothelial cells, several critical questions remain unanswered. What are the precise transcriptional mechanisms that regulate cytokine-induced miRNA expression in brain endothelial cells, and how do different transcription factors (NF-κB, AP-1, and STATs) coordinate to produce specific miRNA expression patterns? How do miRNA networks interact with other regulatory layers, including lncRNAs and circRNAs, to fine-tune BBB responses to inflammatory challenges? What is the extent of heterogeneity in miRNA responses among different brain endothelial cell populations, and how does this heterogeneity contribute to regional differences in BBB vulnerability? Can we identify optimal therapeutic windows for miRNA-based interventions in different disease contexts, and what are the most effective delivery strategies to target brain endothelial cells in vivo while minimizing off-target effects? How do species differences in miRNA sequences, expression patterns, and target genes affect the translatability of findings from animal models to human disease? Finally, can circulating miRNA profiles serve as reliable biomarkers for BBB dysfunction, and can they guide personalized therapeutic interventions in patients with CNS infections or neuroinflammatory diseases?

9 Conclusion

The blood-brain barrier represents a critical interface between the peripheral immune system and the CNS, and its dysfunction during infection and neuroinflammation contributes to significant neurological morbidity and mortality. This review has synthesized current evidence demonstrating that pro-inflammatory cytokines TNF-α and IFN-γ induce specific miRNA changes in human brain endothelial cells that mediate BBB disruption through multiple mechanisms. The identification of barrier-disrupting miRNAs (miR-155, miR-210, miR-143, miR-15a/16-1, miR-494-3p, and miR-212/132) and barrier-protective miRNAs (miR-98, let-7, miR-146a, miR-126, miR-29a/b, miR-27a-3p, miR-130a, miR-150, miR-199a-5p, and miR-132/212) provides a foundation for developing targeted therapeutic strategies.

The complex interplay between cytokines, miRNAs, signaling pathways, and cellular responses highlights the need for systems-level approaches to understand BBB regulation. Feedback regulatory mechanisms involving miRNAs attempt to maintain homeostasis during inflammatory challenges, but can be overwhelmed in severe or chronic inflammation. The context-dependent nature of cytokine effects, influenced by factors such as concentration, duration, and disease-specific mechanisms, underscores the importance of tailored therapeutic approaches for different clinical scenarios.

Lessons from cerebral malaria demonstrate the translational potential of miRNA-based therapies for preserving BBB integrity. The ability to inhibit barrier-disrupting miRNAs or deliver barrier-protective miRNAs using exosomes or other delivery systems offers new therapeutic avenues. However, significant challenges remain in terms of delivery, specificity, timing, and translation from preclinical models to clinical applications.

Future research integrating miRNA biology with other aspects of neurovascular unit function, employing advanced technologies such as single-cell sequencing and systems biology modeling, and conducting rigorous preclinical testing of miRNA therapeutics will be essential for realizing the therapeutic potential of this field. Ultimately, a deeper understanding of cytokine-induced miRNA dysregulation in brain endothelial cells will enable the development of precision medicine approaches to preserve BBB integrity and improve patient outcomes for CNS infections and neuroinflammatory diseases.

Author's note

One or more of the authors of this paper self-identify as an underrepresented ethnic minority in their field of research or within their geographical location.

Generative AI statement

The author(s) declared that generative AI was used in the creation of this manuscript. ChatGPT to correct language and Figurelabs for figures.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Footnotes

Edited by: Sam C. Wassmer, University of London, United Kingdom

Reviewed by: Himanshu Gupta, GLA University, India