Role of Transcription Factors, Noncoding RNAs, Epitranscriptomics and Epigenetics in Post-Ischemic Neuroinflammation

Abstract

Post-stroke neuroinflammation is pivotal in brain repair, yet persistent inflammation can aggravate ischemic brain damage and hamper recovery. Following stroke, specific molecules released from brain cells attract and activate central and peripheral immune cells. These immune cells subsequently release diverse inflammatory molecules within the ischemic brain, initiating a sequence of events, including activation of transcription factors in different brain cell types that modulate gene expression and influence outcomes, the interactive action of various noncoding RNAs (ncRNAs) to regulate multiple biological processes including inflammation, epitranscriptomic RNA modification that controls RNA processing, stability, and translation and epigenetic changes including DNA methylation, hydroxymethylation, and histone modifications crucial in managing the genic response to stroke. Interactions among these events further affect post-stroke inflammation and shape the depth of ischemic brain damage and functional outcomes. We highlighted these aspects of neuroinflammation in this review and postulate that deciphering these mechanisms is pivotal for identifying therapeutic targets to alleviate post-stroke dysfunction and enhance recovery.

Graphical Abstract

Transcription factors, noncoding RNAs, and epitranscriptomic and epigenetic modifications, collectively orchestrate the post-stroke inflammation that plays a significant role in secondary brain damage and functional outcome.

Introduction

Stroke remains the major cause of mortality and long-term disability worldwide, with an annual direct cost of $34 billion for post-stroke care in the United States (Pacheco‐Barrios et al. 2022). As a global health concern, ischemic stroke accounts for ~85% of overall stroke incidents, and the United Nations lists ischemic stroke as a priority target for reducing the burden of non-communicable diseases (Ding et al. 2022). Recombinant tissue plasminogen activator (tPA) and mechanical thrombectomy are the only FDA-approved therapies to treat stroke. However, the use of tPA is limited to a few patients due to the possibility of hemorrhagic transformation and its narrow therapeutic window (within 5h of onset of symptoms) (Du et al. 2021; Davis & Donnan 2009; dela Peña et al. 2017). This emphasizes the significance of continued investigation to understand the molecular mechanisms of post-stroke pathophysiology and identify novel targets to develop treatment strategies to minimize post-stroke brain damage and promote recovery. Many pathophysiological mechanisms triggered after stroke, such as excitotoxicity, ionic imbalance, edema, oxidative stress, endoplasmic reticulum (ER) stress, cell death processes (apoptosis, necrosis and autophagy) and inflammation that occur over varying time frames from minutes to hours to days to precipitate the secondary brain damage (Arruri & Vemuganti 2022; Kim et al. 2019; Morris-Blanco et al. 2022a; Mehta et al. 2007; Balog et al. 2016; Tureyen et al. 2011). Of these, inflammation and its related immune responses play a pivotal role as they can potentiate other pathologic changes like oxidative stress and ER stress (Nakka et al. 2008; Vemuganti & Arumugam 2021). Hence, inflammation is thought to be a critical target for therapeutic intervention to improve stroke outcomes (Elkind 2010; Lambertsen et al. 2019). Armed by both innate and adaptive immune responses, inflammation influences distinct cell types, matrix components, extra- and intracellular receptors, and associated signaling events in the brain during acute, subacute, and chronic stages after stroke (Iadecola & Anrather 2011; Rust et al. 2018; Simats & Liesz 2022). The vascular inflammatory milieu starts due to decreased blood supply and shear stress on endothelium and platelets, exposing the adhesion molecule P-selectin on the endothelial surface that attracts circulating leukocytes (Anrather & Iadecola 2016). Additionally, P-selectin on platelets binds to leukocytes to form intravascular clogs that further obstruct the blood flow and amplify cerebrovascular parenchymal damage (Anrather & Iadecola 2016). Activated endothelial cells also release other adhesion molecules like E-selectin, vascular cell adhesion molecule 1 (VCAM-1), and intercellular adhesion molecule 1 (ICAM-1), which further promote leukocyte adhesion and diapedesis (Andjelkovic et al. 2019; Huang et al. 2006).

On the other hand, the immune response in brain parenchyma is initiated by the release of damage-associated molecular patterns (DAMPs) such as high mobility group box 1 (HMGB1), adenosine triphosphate (ATP), heat shock proteins (HSPs), DNA and RNA from the dying cells, which are then detected by the effectors of the immune system including brain resident microglia via pattern recognition receptors (Gülke et al. 2018). This recognition activates microglia to release inflammatory mediators, including tumor necrosis factor- α (TNF-α), interleukins (ILs) like IL-23, IL-12, and IL-1β, inducible nitric oxide synthase (iNOS), complement proteins, and matrix metalloproteases (MMPs) leading to compromised integrity of the blood-brain barrier (BBB). This allows peripheral immune cells, such as macrophages, neutrophils, T-cells, and lymphocytes, to infiltrate the injury site (Yenari et al. 2010; Simats & Liesz 2022). Ischemic injury-associated DAMPs also induce reactive astrogliosis, leading to further release of inflammatory cytokines such as TNF-α, IL-1α, and interferon-γ (IFN-γ) that aggravate neuroinflammation and subsequent neuronal death (Li et al. 2022b).

Although the inflammatory response after stroke aims to restore cerebral homeostasis, the unregulated acute pro-inflammatory stage provokes rapid loss of neurons, leading to poor functional outcomes. Notably, the loss of function of innate immune cells such as microglia further contributes to increased neuronal loss and brain damage after stroke, suggesting the dual role of inflammation in stroke (Jin et al. 2017). Temporal transcription of pro- and anti-inflammatory genes in the ischemic brain is tightly regulated by transcription factors, noncoding RNAs (ncRNAs), and epigenetic remodelers (Morris-Blanco et al. 2022a; Mehta et al. 2021a). For instance, the transcription factor CCAAT/enhancer binding protein (C/EBP) β, recognized as one of the master regulators of the immune system, is triggered in the subset of neuronal cells within a few hours following an ischemic stroke. This leads to the induction of crucial inflammatory cytokines such as IL-1β and TNF-α (Wu et al. 2023; Kapadia et al. 2006). Another example is the induction of the pro-inflammatory transcription factor early growth response-1 (Egr1) after focal ischemia which is known to induce inflammatory gene expression and secondary brain damage (Tureyen et al. 2008). Once formed, the fate of these inflammatory transcripts is further controlled by the interaction of ncRNAs and epitranscriptomic modifications, suggesting the involvement of multiple layers of inflammatory gene regulation in the ischemic brain (Chokkalla et al. 2022a). Hence, therapies aimed at temporal regulation of inflammatory cascade after stroke may provide better outcomes. This review outlines post-stroke inflammation as a functionally complex response of the immune cells orchestrated by several transcription factors and regulated at the molecular level by ncRNAs, epitranscriptomic and epigenetic changes (Fig. 1).

Figure 1:

Regulation of post-stroke neuroinflammation. Epigenetics and epitranscriptomics and their complex interaction with transcription factors and noncoding RNAs, collectively called epiregulome, modulate gene expression in various cell types in the post-stroke brain. These changes compromise the integrity of the neurovascular unit and BBB, resulting in the infiltration and activation of immune cells and the production of molecules involved in the post-stroke inflammatory sequelae, starting early after the onset of stroke and persisting for days to weeks during the progression of post-stroke brain damage. Me-methylation, Hmc- hydroxymethylation and Ac-acetylation.

Transcription factors control post-stroke neuroinflammation

Regulation of various biological processes, such as inflammation in the brain, depends profoundly on gene expression changes in response to stress. Numerous studies have demonstrated the induction of multiple transcription factors, including hypoxia-inducible factor-1 (HIF-1), cAMP response element-binding protein (CREB), c-fos, peroxisome proliferator-activated receptors (PPAR-α, and γ), p53, IRF-1, activating transcription factor-2 (ATF-2), signal transducer and activator of transcription (STAT) isoforms, nuclear factor kappa B (NF-κB), Egr1, C/EBP α and β, JunD, GLI family zinc finger 2 (Gli2), Sp3, transcription factor AP-2α and spleen focus forming virus proviral integration oncogene (Spi1) that modulate inflammation following ischemic stroke (Kapadia et al. 2006; Satriotomo et al. 2006; Tureyen et al. 2007; Vemuganti 2008; Yi et al. 2007; Tureyen et al. 2008; Zhang et al. 2019b). In particular, IRF-1, NF-κB, ATF-2, STAT3, Egr1, and C/EBPβ could enhance post-ischemic inflammation. In contrast, HIF-1, CREB, c-fos, PPAR-α, PPAR-γ, and p53 have the potential to inhibit post-ischemic inflammation to reduce neuronal damage (Yi et al. 2007). A recent study also showed that JunD, Stat5a/b, and Fos activation prevent ischemic neuronal damage (Kamme & Wieloch 1996; Sola et al. 2005; Yi et al. 2007). In contrast, Cebpb, Egr1, and Rela induction promotes inflammation and leads to neuronal death after cerebral ischemia (Zhang et al. 2019b; Yi et al. 2007). The activity of these transcription factors is tightly regulated to ensure a stable genetic response to injury. We will further highlight some transcriptional factors, such as NF-κB, AP-1, IRFs, and STATs associated with post-stroke inflammation.

Following an ischemic event, the release of various molecules, including reactive oxygen species (ROS), DAMPs, cytokines, and growth factors from the injured neurons, trigger the activation of the pleiotropic transcription factor NF-κB that induces the expression of several pro-inflammatory cytokines/chemokines such as IL-1β, TNF-α, and granulocyte-macrophage colony-stimulating factor, IL-8, macrophage inflammatory protein-1α, and monocyte chemoattractant protein-1 (MCP-1), as well as other molecules such as TNF-receptor associate factor (TRAF), and the Bcl-2 family proteins to promote cell death (Jover-Mengual et al. 2021; Yi et al. 2007). Some of these molecules attract inflammatory cells to enter the brain parenchyma. In addition, microglia activated by NF-κB release inflammatory molecules such as TNF-α, IL-1β, and ROS (Block et al. 2007). Moreover, astrocytic NF-κB triggered by pro-inflammatory stimuli such as TNF-α, IL-1β, IL-17, ROS, and Toll-like receptors promotes the expression of chemokines that attract leukocyte infiltration, resulting in further inflammation in the post-ischemic brain (Linnerbauer et al. 2020). Hence, NF-κB can be a potential target for controlling post-stroke inflammation.

AP-1, a dimeric transcription factor composed of Jun, Fos DNA-binding complexes and activating transcription factors/CREB protein families, regulates cell survival, proliferation, and inflammation. When activated in response to oxidative stress induced by ischemia, AP-1 promotes the expression of pro-inflammatory genes such as IL-1β and TNF-α leading to enhanced platelet aggregation through the RIP1/RIP3/AKT pathway that exacerbates the ischemic brain damage (Diaz-Canestro et al. 2019; Yi et al. 2007; Li et al. 2022c). However, AP-1 can be a double-edged sword as it was also shown to regulate the induction of IL-10 by α-ketoglutarate that protects the post-ischemic brain by inhibiting inflammation via the c-Fos/IL-10/stat3 pathway (Hua et al. 2022).

IRF-1, a nuclear transacting transcription factor responsible for the expression of IFN-α, IFN-β, and other IFN-inducible genes, is known to be induced in the post-stroke brain (Alexander et al. 2003). IRF-1 regulates inflammation and apoptosis, leading to ischemic brain injury (Iadecola et al. 1999). IRF-1 deletion reduced brain damage and improved neurologic recovery in mice subjected to focal cerebral ischemia (Iadecola et al. 1999). Both neurons and glia express various isoforms of the STAT family of transcription factors, including STAT1, STAT2, STAT3, STAT4, STAT5a, STAT5b, and STAT6 (Dziennis & Alkayed 2008). Previous studies also showed the induction of various members of the STAT family of transcription factors in the post-ischemic brain. IL-6, known to be induced after stroke, is a known promoter of neuroinflammation. IL-6 promotes phosphorylation of its receptor-associated Janus kinases (JAKs) like JAK2 that phosphorylates/activates STAT family members like STAT3. When activated, STATs promote IL-6 transcription. This cycle of IL-6-JAK-STAT signaling potentiates inflammation after stroke. Following transient focal ischemia, both pJAK2 and pSTAT3 were reported to be predominantly localized in the macrophages/microglia (Satriotomo et al. 2006). Importantly, either inhibition of JAK2 or knockdown of STAT3 led to significant neuroprotection and better functional recovery in rodents subjected to transient focal ischemia (Satriotomo et al. 2006). Localized STAT3 activation in microglia by homocysteine following ischemic stroke was also shown to exaggerate neuroinflammation (Chen et al. 2017). Furthermore, blocking STAT3 signaling by nuclear localization of the transcriptional cofactor Yes-associated protein (YAP) in astrocytic cells protects the brain against ischemic damage (Huang et al. 2020). These studies indicate a derogatory role for STAT3 following cerebral ischemia. However, some studies also showed the beneficial nature of STAT3 after stroke. For instance, estradiol-mediated neuroprotection against ischemic brain injury was facilitated by STAT3 activation and abolished by its inhibition (Dziennis et al. 2007). Thus, the role of STATs might be distinct based on the type and severity of the brain injury.

Besides controlling the induction of inflammatory molecules, transcription factors also participate in the inflammatory cascade by regulating the expression of ncRNAs. Various transcription factors including c-Myc, p53, NF-κB, and HIF-1α associated with the redox status are responsible for miRNA biogenesis and downstream changes (Carbonell & Gomes 2020). For example, transcription factor p53 regulates miR-34a and miR-145 and thereby influences the microglial behavior (Su et al. 2014). Similarly, NF-κB activation modulates the expression of various pro-inflammatory genes and miRNAs such as miR-9, miR-21 miR-146a, and miR-155 which are known to target inflammatory mRNAs (Xue et al. 2018; Zhan et al. 2023; Adly Sadik et al. 2021; Qu et al. 2019). Similarly, transcription factor E2F1 can directly modulate the transcription of miR-122 following ischemic stroke (Wu et al. 2020). Furthermore, changes in lncRNAs and circRNAs following stroke are regulated by a set of transcription factors (Mehta et al. 2017; Cao et al. 2020; Mehta et al. 2023a). For instance, we recently reported that NF-κB aggravates post-stroke brain damage by promoting the expression of pro-inflammatory lncRNA FosDT (Mehta et al. 2023a).

In addition, the crosstalk among transcription factors, their epigenetic regulation, and accessibility to DNA binding sites could influence the post-stroke gene expression and, thus, pathologic processes such as inflammation that modulate neuronal survival and/or damage. Understanding these diverse roles of transcription factors in the post-stroke brain is crucial for developing targeted therapies.

Noncoding RNAs and ischemic brain damage

The majority of the mammalian transcriptome is ncRNAs, which are pervasively transcribed from the genome (Beermann et al. 2016). Although they lack protein-coding ability, ncRNAs regulate gene expression, chromatin remodeling, RNA processing, and transgenerational epigenetics (Statello et al. 2021; Fitz-James & Cavalli 2022). Moreover, pathological implications of ncRNAs in several human diseases underscore their therapeutic potential (Esteller 2011). Different types of ncRNAs have been discovered in addition to ribosomal RNAs (rRNAs) and transfer RNAs (tRNAs) which do not encode proteins but are involved in the process of protein synthesis. The list of ncRNAs is still expanding, with the extensively studied ones such as microRNAs (miRNAs), long noncoding RNAs (lncRNAs), and circular RNAs (circRNAs) to relatively less explored types like PIWI-interacting RNAs (piRNAs), small nuclear RNAs (snRNAs), small nucleolar RNAs (snoRNAs), large intergenic ncRNAs (lincRNAs), and enhancer RNAs (eRNAs) (Esteller 2011; Sartorelli & Lauberth 2020). We and others previously showed the differential expression of several miRNAs, piRNAs, lncRNAs, and circRNAs in the post-stroke brain and revealed the mechanisms by which specific ncRNAs modulate the ischemic pathophysiology (Dharap et al. 2009; Mehta et al. 2015; Mehta et al. 2023b; Mehta et al. 2023c; Zhang et al. 2019a). Accumulating evidence also suggests the role of these ncRNAs in regulating post-stroke neuroinflammation (Lu et al. 2020a). The following sections discuss the experimental evidence linking specific ncRNA subtypes to post-stroke cerebral inflammation (Fig. 2 and Table 1).

Figure 2:

Noncoding RNAs and their interaction modulate the post-stroke inflammatory response and secondary brain damage. Various miRNAs (including miR-143, 181a, 193a, 381–3p, 424 and 671), lncRNAs (such as ANRIL, H19, HCG11, NEAT1 and Tug1) and circRNAs (like CDC14A and DLGAP4) altered in various central and peripheral immune and non-immune cell types modulate post-stroke inflammation independently as well as in concert in a spatiotemporal manner, and thereby regulates post-stroke secondary brain damage.

Table 1:

Role of ncRNAs in post-stroke inflammation

Model	ncRNA	Mechanism	Outcome
MCAO; C57BL/6 mice (Han et al. 2023)	miR-193a↓	Represses ubiquitin-conjugating enzyme V2 thereby promoting neutrophil reprogramming towards an anti-inflammatory N2 phenotype	AgomiR treatment decreased infarct volume and improved motor function
MCAO; C57BL/6 mice (Zhao et al. 2013)	miR-424↓	Regulate neuronal apoptosis and microglial activation by repressing G1 phase cell cycle activators like CDC25A, cyclin D1, and CDK6	Mimic decreased infarct volume and brain edema
MCAO; C57BL/6 mice (Kolosowska et al. 2020)	miR-669c-3p↑	Inhibits inflammation by repressing MyD88	Lentiviral-mediated overexpression decreased infarct volume and improved neurologic function
MCAO; C57BL/6 mice (Huang et al. 2018)	miR-210↑	Regulates pro-inflammatory mediators such as TNF-α, IL-1β, IL-6, CCL2, and CCL3	miR-210-LNA reduced infarct volume and improved motor function
MCAO; C57BL/6 mice (Yang et al. 2017)	miR-15a/16–1↑	Controls expression of pro-inflammatory cytokines including, IL-6, MCP-1, VCAM-1, and TNF-α	Antagomir treatment decreased infarct volume and improved motor function
MCAO; C57BL/6 mice (Xu et al. 2015)	miR-181a↑	Modulates anti-apoptotic proteins Bcl-2, XIAP expression and NF-κB activation	Antagomir administration reduced infarct volume and improved neurologic function
MCAO; SHR & SD rats; C57BL/6 & db/db mice (Mehta et al. 2021a; Mehta et al. 2015; Mehta et al. 2023a)	lncRNA FosDT↑	Regulates the expression of genes involved in inflammatory response, neutrophil chemotaxis, & cytokine production	Inhibition or deletion decreased infarct volume and improved motor and cognitive function
MCAO; C57BL/6 mice; OGD/R in BV2 cells (Wang et al. 2017)	lncRNA H19↑	Regulates TNF-α, IL-1β and IL-10 and reduced microglial activation; Induces HDAC to suppress M2 polarization	Inhibition reduced infarct volume, brain edema, and improved motor function
OGD/R in BV2 cells (Ni et al. 2020)	lncRNA NEAT1↑	Modulates M1 markers CD16, CD32 and CD86	Inhibition decreased microglial activation
MCAO; C57BL/6 mice; OGD/R in BMECs and N2a cells (Zhang et al. 2017).	lncRNA MALAT1↑	Controls the expression of inflammatory markers E-selectin, MCP-1 and IL-6; anti-inflammatory activity of MALAT1 may be due to its direct binding with E-selectin	Deletion increased infarct volume and impaired motor function
MCAO; C57BL/6 mice; OGD/R in BV2 cells (Li et al. 2020)	lncRNA MEG3↑	Inhibits KLF-4 expression to prevent microglial M2 polarization	Knockdown inhibited M1 microglial polarization and inflammation
MCAO; C57BL/6 mice (Mehta et al. 2023b)	circRNA CDR1as↓	Alters the expression of IL-1β	Overexpression decreased infarct volume and improved motor function
MCAO; C57BL/6 mice (Zuo et al. 2021)	circRNA CDC14A↑	Regulates astrocyte activation and neutrophil N2 polarization	Inhibition decreased infarct volume and improved neurologic function

Bcl-2, B-cell lymphoma-2; BMEC, brain meningeal endothelial cells; CCL2, chemokine ligand 2; CCL3, chemokine ligand 3; CD, cluster of differentiation; CDC25A, cell division cycle 25 A; CDK6, cyclin-dependent kinase 6; db/db, diabetic; HDAC1, histone deacetylase 1; IL, interleukin; KLF-4, krüppel-like factor 4; MCP-1, monocyte chemoattractant protein-1; MyD88, myeloid differentiation primary response-88; NF-κB, nuclear factor kappa-light-chain-enhancer of activated B cells; OGD/R, oxygen-glucose deprivation/reperfusion; MACO, middle cerebral artery occlusion; SHR, spontaneously hypertensive; TNF-α, tumor necrosis factor-alpha; VCAM-1, vascular cell adhesion molecule-1; XIAP, X-linked inhibitor of apoptosis protein.

MiRNAs and post-stroke neuroinflammation

MiRNAs are small evolutionarily conserved ncRNAs of 21 to 24 nucleotides in length that play a critical role in post-transcriptional regulation of gene expression either by accelerating mRNA degradation or reducing the translation efficiency (Bartel 2004). Most miRNAs are coded from intronic or intergenic regions, while some are also processed from the exons of long non-protein coding transcripts (Rodriguez et al. 2004). The miRNA loci of intronic regions of coding or noncoding transcription units are transcribed by RNA polymerase II (Pol II) as part of their host gene transcription, whereas the intergenic miRNAs possess their promoters and are transcribed as independent transcription units by Pol II (Kim & Kim 2007; Lin et al. 2006). Biogenesis of miRNAs starts in the nucleus, resulting in a long primary transcript with a stem-loop structure containing miRNA sequence, known as primary miRNA (pri-miRNA), that undergoes a series of maturation steps (Lee et al. 2002). The nuclear RNase III Drosha then trims a pri-miRNA to release a pre-miRNA, which is exported to the cytosol by exportin 5 and cleaved by Dicer to generate a small RNA duplex (Bartel 2004). This small RNA duplex is cleaved by argonaute into a guide and passenger strand. The guide strand is loaded into an effector complex RNA-induced silencing complex (RISC) to regulate its target mRNAs by binding to their 3’-untranslated region (UTR) (Bartel 2004). miRNA biogenesis is tightly controlled at multiple levels, from pri-miRNA transcription to mature miRNA modifications, including RNA methylation, adenylation, and uridylation (Ha & Kim 2014). Dysregulated expression of miRNAs is linked to the pathogenesis of several diseases, including those of neurological origin. In ischemic stroke, initial findings from our group and others revealed the differential expression of cerebral miRNAome after focal ischemia in rats (Dharap et al. 2009; Jeyaseelan et al. 2008). Accumulating evidence suggests the involvement of miRNAs in the etiology of secondary brain damage after stroke (Li et al. 2018).

MiRNAs regulate both early and delayed phases of inflammatory signaling following ischemic injury to the brain (Khoshnam et al. 2017). Innate immune cells, including neutrophils, monocytes, macrophages, and natural killer cells, that start infiltrating within hours after ischemic stroke onset will release many pro-inflammatory molecules that promote inflammation (Iadecola et al. 2020). For example, miR-193a levels were reported to be decreased in neutrophils of acute ischemic stroke patients compared to healthy individuals (Han et al. 2022). Mechanistically, miR-193a promotes neutrophil reprogramming towards an anti-inflammatory N2 phenotype by inhibiting ubiquitin-conjugating enzyme V2 (Han et al. 2022). In acute ischemic stroke patients treated with tPA, increased levels of miR-193a correlated well with the improved outcomes at three months of the follow-up period, evaluated using a Modified Ranking Scale (Han et al. 2022). Similarly, miR-193a agomiR administration decreased infarct volume and improved motor function after transient focal ischemia in mice subjected to middle cerebral artery occlusion, suggesting the beneficial effects of miR-193a (Han et al. 2022).

Acute ischemic stroke patients showed lower plasma levels of miR-424, which correlated with stroke outcomes (Zhao et al. 2013). Mice subjected to permanent cerebral ischemia also showed decreased levels of miR-424 in plasma and ipsilateral brain tissue, and treatment with a miR-424 mimic attenuated post-stroke microglial activation and TNF-α levels leading to significant neuroprotection (Zhao et al. 2013). miR-669c-3p levels were increased in the peri-infarct cortex of mice after permanent or transient cerebral ischemia (Kolosowska et al. 2020). Lentiviral-mediated overexpression of miR-669c-3p decreased brain damage and improved neurological function recovery in mice subjected to transient focal ischemia (Kolosowska et al. 2020). This beneficial effect of miR-669c-3p was observed to be associated with increased expression of microglial arginase 1, chitinase-like 3, and PPAR-γ, and inhibition of myeloid differentiation primary response-88 (MyD88) signaling leading to attenuation of the microglial pro-inflammatory response (Kolosowska et al. 2020). Following transient focal ischemia in adult rats, cerebral miR-210 levels were upregulated and remained elevated for up to 7 days, and its inhibition with the LNA-oligonucleotides suppressed the release of pro-inflammatory mediators such as TNF-α, IL-1β, IL-6, CCL2, and CCL3 (Jeyaseelan et al. 2008; Huang et al. 2018; Lou et al. 2012). Further, the inhibition of miR-210 with the LNA-oligonucleotides injected 24 hours prior or 4 hours after the induction of ischemia decreased infarct volume and improved motor function in rodents (Huang et al. 2018). The highly conserved miRNAs miR-15a and miR-16–1 function concertedly as a cluster to repress their shared mRNA targets (Calin et al. 2008). Their levels were found to be elevated in the plasma of acute ischemic stroke patients (Tan et al. 2009). The miR-15a/16–1 knockout mice showed resilience to ischemic injury when subjected to transient focal ischemia (Yang et al. 2017). Furthermore, antagomiR-15a/16–1 decreased brain damage and improved motor function in mice subjected to transient focal ischemia (Yang et al. 2017). Mechanistically, miR-15a/16–1 inhibition curtailed the expression of pro-inflammatory cytokines, including IL-6, MCP-1, VCAM-1, and TNF-α (Yang et al. 2017). In mice, post-ischemic induction of miR-181a was thought to be a promoter of secondary brain damage, and its inhibition reduced the infarct volume and improved long-term motor function (Xu et al. 2015). Neuroprotective effects of miR-181a inhibition was partly mediated through suppression of the NF-κB activation (Xu et al. 2015). Thus, several miRNAs were shown to have the potential to be the regulators of post-stroke neuroinflammation and the associated secondary brain damage.

LncRNAs play a role in post-stroke neuroinflammation

LncRNAs are noncoding transcripts of >200 nucleotides in length (Mercer et al. 2009). To date, ~60,000 lncRNAs have been identified in humans (Derrien et al. 2012; Iyer et al. 2015). Most lncRNAs are transcribed by Pol II and possess m⁷G caps at 5′-end and poly(A) tails at 3′-end, similar to mRNAs (Quinn & Chang 2016). However, cis-regulatory capacity, lack of open reading frames, unique 3’-end processing, ability to act as templates for nucleic acid polymerization, and molecular scaffolding capability to recruit various chromatin-modifying proteins distinguish lncRNAs from mRNAs (Quinn & Chang 2016). LncRNAs regulate gene expression at nearly all stages through imprinting genomic loci, epigenetic silencing, controlling chromatin structure and function, and affecting RNA splicing, stability, and translational efficiency (Statello et al. 2021); hence, their dysregulation contributes to pathological changes associated with several diseases including stroke (Statello et al. 2021; Chen et al. 2021).

Our lab showed that Fos downstream transcript (FosDT), a highly conserved brain-enriched and stroke-responsive lncRNA, promotes brain damage in rats after transient cerebral ischemia (Mehta et al. 2015). FosDT knockout or knockdown decreased the infarct size and improved motor function recovery after stroke compared to their wild-type/control cohorts (Mehta et al. 2021a; Mehta et al. 2015). This action of FosDT is thought to be partly mediated by scaffolding chromatin modifying protein and regulating the induction of stroke-responsive inflammatory genes (Mehta et al. 2021a; Mehta et al. 2015).

Epigenetic reprogramming of microglial phenotype regulates neuroinflammation and contributes to secondary brain damage after ischemic stroke, which may be regulated by lncRNA H19 (Patnala et al. 2017). H19 was observed to be elevated in the blood of stroke patients compared to the healthy controls (Wang et al. 2017). Additionally, there was a positive correlation between plasma H19 levels (within 3 hours of onset) and the NIH Stroke Scale (NIHSS) score at admission and 7 days after thrombolytic therapy in stroke patients (Wang et al. 2017). A similar trend was noticed between neutrophilic H19 levels and plasma TNF-α levels in stroke patients (Wang et al. 2017). Moreover, a significant increase in H19 levels was also observed in the plasma, white blood cells, and brain of mice after transient focal ischemia. Interestingly, H19 knockdown is beneficial by decreasing infarct volume and edema and improving motor function after transient focal ischemia (Wang et al. 2017). Mechanistically, H19 knockdown curtailed TNF-α and IL-1β levels and augmented IL-10 expression, indicating the immunomodulatory role of H19 after stroke (Wang et al. 2017). Additionally, H19 knockdown shifts BV2 microglial phenotype from pro-inflammatory to anti-inflammatory by repressing histone deacetylase 1 (HDAC1) upon oxygen-glucose deprivation injury (Wang et al. 2017). Although the direct mechanistic link between H19 and HDAC1-mediated microglial reprogramming is yet to be determined, a recent study demonstrated that H19 sponges miR-29b to derepress its target C1QTNF6 in leukocytes, thereby promoting the release of IL-1β and TNF-α which increases post-stroke neuroinflammation after transient focal ischemia in rodents (Li et al. 2022a). Microglial polarization is also influenced by lncRNA NEAT1, which was shown to be elevated in the blood of acute ischemic stroke patients compared to healthy individuals. NEAT1 levels also positively correlated to NIHSS scores and infarct size (Ni et al. 2020). It prevents BV2 microglial polarization towards pro-inflammatory phenotype after ischemic conditions (Ni et al. 2020).

Plasma levels of lncRNA MALAT1 were significantly low in acute ischemic stroke patients compared to healthy individuals, and its expression negatively correlated to NIHSS score and the levels of inflammatory mediators such as TNF-α, IL-6, IL-8, and IL-22 (Ren et al. 2020). However, MALAT1 was shown to be upregulated in the cerebral microvessels of mice after transient cerebral ischemia, and MALAT1 knockout mice showed increased brain damage, neurological deficit, and compromised motor function compared to their wild-type counterparts (Zhang et al. 2017). Mechanistically, MALAT1 knockout mice showed increased cortical expression of inflammatory mediators such as IL-6, E-selectin, and MCP-1 after ischemic stroke, indicating the anti-inflammatory activity of MALAT-1 in response to cerebral ischemia (Zhang et al. 2017). The lncRNA MEG3 was shown to be upregulated and its knockdown promoted better neurological functional recovery after transient cerebral ischemia in mice (Yan et al. 2016; Li et al. 2020). The beneficial effect of MEG3 knockdown was attributed to a shift in microglial polarization towards anti-inflammatory phenotype along with a reduction in the expression of pro-inflammatory cytokines such as TNF-α and IL-1β (Li et al. 2020). These studies indicate the role of lncRNAs in regulating post-stroke inflammatory response and indicate their potential as targets for stroke therapy.

CircRNAs and post-stroke neuroinflammation

CircRNAs are covalently closed, single-stranded ncRNAs formed by the back-splicing of precursor mRNA between a downstream 3’ splice site and an upstream 5’ splice site (Chen 2016). CircRNAs are evolutionarily conserved, widely expressed, and tightly regulated in various physiological contexts with significant potential to regulate gene expression by sponging miRNAs (Chen 2020) (Patop et al. 2019). Accumulating evidence also suggests that circRNAs can control protein stability as well as act as molecular scaffolds to facilitate protein activity (Santer et al. 2019). Furthermore, certain circRNAs can regulate their host gene transcription in the nucleus by interacting with U1 snRNAP (Li et al. 2015). An intriguing feature of circRNAs is their resistance to RNase R-mediated degradation, which makes them potential targets for therapeutic interventions (He et al. 2021; Mehta et al. 2020).

CircRNAs are highly enriched in the mammalian brain, and their abundance in the synapses is high during neuronal differentiation. In the synapse, they are expressed more than their mRNA isoforms, suggesting their essential role in neuronal signal transmission (Rybak-Wolf et al. 2015). We observed that hundreds of circRNAs are differentially and temporally altered in the peri-infarct cortex of mice following transient cerebral ischemia (Mehta et al. 2017). Of the differentially expressed circRNAs, CDR1as (ciRS-7) was downregulated over an extended reperfusion time point after transient focal ischemia (Mehta et al. 2017; Mehta et al. 2023b). We further showed that overexpression of CDR1as conferred resilience to ischemic injury by increasing miR-7 abundance and thereby repressing α-synuclein. miR-7 is a prosurvival miRNA essential for post-stroke functional recovery (Mehta et al. 2023c). Its regulation by CDR1as is thus crucial for curtailing post-stroke brain damage in mice (Mehta et al. 2023b). Neuroprotective effects of CDR1as are mediated partly by reducing IL-1β release from activated microglia to curtail post-stroke neuroinflammation (Mehta et al. 2023b). Although the direct mechanistic link between CDR1as and inflammation is not established yet, a recent study showed that overexpression of CDR1as promotes switching of macrophages to anti-inflammatory phenotype upon challenging with IFN-γ and TNF-α, suggesting the possible anti-inflammatory potential of CDR1as (Gonzalez et al. 2022). CircRNA circHECTD1 was upregulated in peripheral blood mononuclear cells of acute ischemic stroke patients compared to healthy controls and positively correlated with NIHSS score and serum levels of pro-inflammatory cytokines (TNF-α, IL-1β, and IL-6) (Peng et al. 2019).

Further, plasma levels of circFUNDC1, circPDS5B, and circCDC14A were significantly higher in acute ischemic stroke patients than in healthy controls, and their levels were positively linked with infarct volume (Zuo et al. 2020). Intriguingly, circPDS5B levels were noted to be higher in lymphocytes, whereas cirCDC14A levels were higher in granulocytes of stroke patients (Zuo et al. 2020). Pathway analysis showed that all 3 of these circRNAs could aggravate inflammatory responses after stroke (Zuo et al. 2020). High levels of circCDC14A were also observed in the neutrophils and peri-ischemic cortical astrocytes of mice after transient cerebral ischemia (Zuo et al. 2021). An increase in circCDC14A might be detrimental to the brain as its knockdown decreased ischemic brain damage and improved neurologic function, mainly by suppressing astrocyte activation and modulating neutrophil polarization (higher proportion of N2 phenotype) after stroke (Zuo et al. 2021). Overall, these evidence points out the role of circRNAs in regulating post-stroke neuroinflammation. Therefore, the immunomodulatory function of circRNAs can be used to design new therapeutic possibilities for treating stroke.

Various classes of noncoding RNAs interact to control post-stroke neuroinflammation

Various classes of ncRNAs, including lncRNAs, circRNAs, and miRNAs, form intricate regulatory networks that influence gene expression and cellular processes and thus play crucial roles in governing both pathological and physiological functions within the brain (Mehta et al. 2021b). The miRNAs bind to 6–8 nucleotide regions called miRNA response elements (MREs) in the 3’-UTRs of mRNAs, thus controlling their translation (Wu et al. 2013). Many ncRNAs like lncRNAs and circRNAs also contain MREs and hence act as either miRNA sponges or decoys to prevent miRNAs from binding and inhibiting their target mRNAs (Mehta et al. 2021b; Yamamura et al. 2018).

Several circRNAs and lncRNAs contain a full or partial complementary of the same MRE and hence can bind to a specific miRNA with different stringencies. These are termed competing endogenous RNAs (ceRNAs) that titrate the action of a miRNA in a complex manner (Salmena et al. 2011). Furthermore, many ncRNAs contain multiple MREs that can concurrently bind to several miRNAs and thus precisely control gene expression patterns critical for brain development, function, as well as response to pathological events such as stroke (Yang et al. 2023).

Role of lncRNA-miRNA interactions in post-stroke inflammation

Many lncRNAs observed to be dysregulated in either stroke patients and/or in animal models of cerebral ischemia can act as ceRNAs to sponge miRNAs and thus control many post-stroke pathologic events, including neuroinflammation (Dykstra-Aiello et al. 2016; He et al. 2018; Dharap et al. 2012; Bhattarai et al. 2017; Zhang et al. 2020). For instance, lncRNA SNHG14 is known to be induced in the post-stroke brain and modulates microglial activation and inflammatory response by sponging miR-136, miR-145, and miR-199b (Qi et al. 2017; Zhong et al. 2019; Zhang et al. 2021). Sequestering of miR-145 by SNHG14 allows the expression of microglial activation gene PLA2G4A and, thus, inflammation after stroke (Qi et al. 2017). Silencing of SNHG14 was shown to protect the ischemic brain by suppressing inflammation as well as reducing edema by regulating miR-199b/miR-145-dependent glial water channel aquaporin-4 (AQP4) after cerebral ischemia (Wang et al. 2020a; Zhang et al. 2021).

In patients with ischemic stroke, coordinated upregulation of lncRNA H19 and an inflammation-related gene, the Complement C1q Tumor Necrosis factor-related Protein 6 (C1QTNF6), and downregulation of miR-29b were observed in neutrophils (Li et al. 2022a). A similar change in H19 and miR- 29b expression was also observed on day one of reperfusion in the brains of rats subjected to transient focal ischemia (Li et al. 2022a). Intriguingly, further downregulation of miR-29b with an antagomir worsened, whereas H19 inhibition alleviated ischemic brain injury potentially by regulating C1QTNF6, BBB disruption, and release of IL‐1β and TNF‐α by leukocytes during ischemic injury (Li et al. 2022a). The potential involvement of H19 in inflammation extends beyond ischemic stroke. A recent study showed that H19 activation following subarachnoid hemorrhage in rats led to down-regulation of miR-138, which in turn derepressed NOD-like receptor pyrin domain containing 3 (NLRP3)-mediated inflammasome activation (Liu et al. 2023). Furthermore, H19 knockdown prevented NLRP3 inflammasome activation after hemorrhagic stroke in rats (Liu et al. 2023). Regulation of NLRP3-mediated inflammasome signaling by a single lncRNA or miRNA might not adequately control the post-stroke inflammation. A recent study showed that lncRNA Tug1 contributes to post-stroke NLRP3 inflammasome-dependent pyroptosis via miR-145a and toll-like receptor 4 (Tlr4) (Yao et al. 2022). Tug1 induced after focal ischemia was shown to sequester miR-145a, leading to derepression of its pro-inflammatory and pro-apoptotic target genes Tlr4, IL-1β, IL-18, NLRP3, ASC, cleaved caspase-1, gasdermin D, and p65 (Yao et al. 2022). Tug1 inhibition with an intraventricular infusion of AAV-shTug1 reversed this sequence, leading to reduced microglia activation, decreased infarct volume, and improved neurological function following experimental stroke in mice (Yao et al. 2022). High levels of lncRNA UCA1 and low levels of miR-18a were reported in the blood of acute ischemic stroke patients (Yan et al. 2023). Silencing UCA1 increased the expression of miR-18a in rats subjected to ischemic stroke and suppressed inflammation (Yan et al. 2023). While the precise mechanism through which miR-18a modulates ischemic brain injury remains unclear, a recent study demonstrated that miR-18a has the potential to attenuate ischemic injury by targeting Tlr8 and subsequently modulating downstream Tlrs/NF-κB signaling pathways known to cause inflammation (Lu et al. 2020b). Furthermore, stopping ischemia-induced lncRNA ANRIL with siRNA has a beneficial effect on reducing neuroinflammation in a middle cerebral artery occlusion stroke model in mice by upregulating miR-671–5p expression and decreasing miR-671 target NF-κB. ANRIL inhibition further decreased NF-κB responsive levels of pro-inflammatory cytokines IL-1β, IL-6 and TNF-α, and iNOS and preserved expression of tight junction proteins (ZO-1, occludin, and claudin 5), thereby alleviating the BBB disruption in mice subjected to transient focal ischemia (Deng et al. 2022) (Deng et al. 2022).

A recent study showed that reduced lncRNA ZFAS1 expression promotes inflammation and neurological impairment in acute ischemic stroke patients (Wang et al. 2022). In a mouse model of ischemic stroke, treatment with bone marrow mesenchymal stem cells (BMSC) derived exosomes carrying ZFAS1 alleviated inflammation by inhibiting miR-15a and lowering the levels of IL-1β, IL-6, and TNF-α (Yang & Chen 2022). Whereas the effect of BMSC-released exosomal ZFAS1 was counteracted when miR-15a was overexpressed, indicating the significant role of ZFAS1 in post-stroke inflammation and brain damage (Yang & Chen 2022). It is unclear if miR-15a directly affects IL-1β, IL-6, and TNF-α levels, thereby affecting inflammation. However, it is suggested that mice with endothelial cell-selective deletion of miR-15a/16–1 cluster that directly binds to the 3′UTR of Claudin-5, showed smaller brain infarcts, reduced BBB leakage and decreased infiltration of peripheral immune cells after transient focal ischemia (Ma et al. 2020). As endothelial claudin-5 is critical for post-stroke BBB disruption, thus inhibiting miR-15a directly or indirectly by overexpressing ZFAS1 could be a novel strategy to reduce the infiltration of immune cells. LncRNA HCG11 was found to be upregulated after transient focal ischemia in rat brains. When silenced, HCG11 inhibited the increase in neuroinflammatory molecules (TNF-α and IFN- γ), restricted the enlargement of cerebral infarction, and improved neurological recovery following cerebral ischemia in rats (Gao et al. 2022). HCG11 targets miR-381–3p in the ischemic brain and negatively regulates its levels (Gao et al. 2022). However, co-inhibiting miR-381–3p with an antagomir and HCG11 with siRNA enhanced ischemic brain injury in rats attenuated by siRNA alone. These effects of HCG11 and miR-381–3p were mediated by p53, a known promotor of secondary ischemic brain damage and a critical regulatory factor in the Wnt signaling and apoptosis (Vaseva et al. 2012; Gao et al. 2023). p53 can produce Wnt ligands that stimulate macrophages to produce IL-1β and drive inflammation (Wellenstein et al. 2019; Gao et al. 2023). Consequently, the interaction between miR-381–3p and the 3’UTR of p53 regulates p53 expression. The suppressive impact of miR-381–3p on p53 can be reversed using antagomir, increasing p53 expression and triggering subsequent downstream alterations (Gao et al. 2022). The findings of all these studies provide evidence that the interplay between lncRNAs and miRNAs modulates mRNA functions and post-stroke inflammation.

CircRNA-miRNA interactions and post-stroke inflammation

CircRNAs serve multiple functions, such as sponging RNA binding proteins (RBPs) and regulating transcription, translation, and splicing. In addition to these, they bind and control miRNAs. This function controls the ceRNA networks (circRNA-miRNA-mRNA) to regulate the translation of target mRNAs. CDR1as, which is highly expressed in human and mouse brains predominantly in excitatory neurons, is an example of a circRNA playing a regulatory role in a ceRNA network (Hansen et al. 2011; Hansen et al. 2013). Bioinformatics showed an unusually high number of conserved binding sites (~70) for miR-7 in CDR1as (Hansen et al. 2013; Memczak et al. 2013). CDR1as pull-down with Ago2 confirmed a high degree of Ago2 occupancy of miR-7 (Hansen et al. 2013). CDR1as deficiency generated using a loss-of-function using CRISPR-Cas9 removal of CDR1as locus caused compromised sensorimotor gating and disrupted synaptic transmission but normal social behavior, unaffected anxiety levels, unperturbed locomotor activity, and recognition memory or exploratory behavior in mice (Piwecka et al. 2017). Intriguingly, the deletion of CDR1as resulted in post-transcriptional deregulation of mature but not precursor miR-7 precisely in the neural tissues but not outside the brains of knockout mice (Piwecka et al. 2017). This further resulted in the upregulation of miR-7 targets, including Nr4a3, Irs2, and Klf4 in the CDR1as knockout mice brains, indicating the significance of circRNA-miRNA interactions in regulating the expression of target mRNAs and maintaining normal brain function (Piwecka et al. 2017).

Accumulating evidence suggests that circRNA/mRNA networks are crucial in controlling various neural functions, including plasticity, synaptogenesis, and neuronal differentiation, and pathological processes such as inflammation (Mehta et al. 2020; Piwecka et al. 2017; Mehta et al. 2023b; Li et al. 2015). Inflammation-related circRNA polymorphisms were shown to be associated with functional outcomes in patients with ischemic stroke (Liu et al. 2021). The circ-STAT3 rs2293152 carrying GG genotype exhibited worse outcomes three months after stroke and can serve as an independent risk factor for stroke recovery (Liu et al. 2021).

The circDLGAP4 functions as an endogenous miR-143 sponge (Bai et al. 2018). CircDLGAP4 was down-regulated and miR-143 was upregulated in both the post-ischemic mouse brain and the plasma of acute stroke patients (Bai et al. 2018). This led to the repression of miR-143 target proteins claudin-5, occludin, and ZO-1 (tight junction proteins that control BBB integrity) in mice subjected to cerebral ischemia (Bai et al. 2018). Overexpressing circDLGAP4 following transient cerebral ischemia in mice increased the levels of tight junction proteins, leading to better cerebrovascular integrity (Bai et al. 2018). The compromised BBB results in disrupted ion balance, edema, and neuroinflammation, which may lead to neuronal dysfunction and neuronal degeneration associated with stroke. The circRNA circTTC3, induced in post-ischemic mouse brains and astrocytes exposed to oxygen-glucose deprivation, targets Tlr4 by sponging miR-372–3p (Yang et al. 2021). This might contribute to post-ischemic inflammation and brain damage, as Tlr4 is a known promoter of inflammation (Caso et al. 2007). Similarly, circ_0000831 is significantly downregulated in mice brain tissues subjected to centrifugation-mediated vertigo a day after transient focal ischemia (Huang et al. 2022). This circRNA is suggested to hamper inflammation as its overexpression with adenovirus injected into the lateral ventricle three days before the induction of ischemia suppressed miR-16, a miRNA that was also detected with high expression in the ischemic brain tissue and peripheral blood of ischemic stroke patients (Kim & Lee 2017; Huang et al. 2022). miR-16 inhibitor resulted in upregulating its target gene AdipoR2 in microglia and promoting PPAR-γ expression (Huang et al. 2022). PPAR-γ induction improves post-stroke neurologic function and reduces brain damage by decreasing TNF-α, IL-1β, IL-6, MMP-2, and MMP-9 in rats (Wu et al. 2018). Intriguingly, simultaneous overexpression of circ_0000831 and silencing of PPAR-γ in stroke mice resulted in higher TNF-α and IL-1β levels and a higher ratio of p-p65/p65 than circ_0000831 overexpression alone, suggesting that circ_0000831 plays an essential role in regulating neuroinflammation via miR-16–5p/AdipoR2/PPARγ axis in the post-stroke brain (Huang et al. 2022). Furthermore, circCDC14A serves as a miR-23a-3p sponge and facilitates the expression of the chemokine stromal-derived factor-1 (CXCL12), which plays a role in the post-ischemic inflammatory response (Huo et al. 2022; Wang et al. 2012). CircCDC14A is induced in mice and cell models of ischemic stroke (Huo et al. 2022). CircCDC14A knockdown enhanced cell survival while inhibiting miR-23a-3p or overexpressing circCDC14A counteracted this in HT22 cells exposed to oxygen-glucose deprivation (Huo et al. 2022). We recently showed that the circRNA CDR1as regulates the abundance of α-synuclein-mediated ischemic brain damage by controlling miR-7 (Mehta et al. 2023b). α-synuclein is a promoter of ischemic brain damage and is also an inducer of neuroinflammation in neurodegenerative diseases (Du et al. 2018; Kim et al. 2016). We observed that increasing miR-7 levels with a mimic suppresses α-synuclein translation, leading to better neuronal survival after focal ischemia in rodents (Kim et al. 2018; Mehta et al. 2023c). The association between CDR1as and miR-7 is intricate. Rather than inhibiting its function, CDR1as may act to protect, stabilize, and transport miR-7 within the cells (Mehta et al. 2023b; Piwecka et al. 2017). These findings indicate that the interactions between circRNAs and miRNAs play a crucial role in modulating post-stroke inflammation.

Inflammation and post-stroke crosstalk of circRNA, miRNA and lncRNA

The complexities of the ceRNA network increase when a lncRNA and circRNA compete for a single target miRNA, and this serves as an extra layer of genetic regulation (Mehta et al. 2021b; Kleaveland et al. 2018). Targeting of a miRNA by ncRNAs might have opposing effects. For example, circRNA CDR1as and lncRNA Cyrano compete for miR-7 in the brain (Kleaveland et al. 2018). Cyrano uses an extensive paired site to bind to the 3’ region of miR-7 that triggers a structural alteration recognized by ZSWIM8 Cullin-RING E3 ubiquitin ligase that polyubiquitinates the Ago protein, leading to its proteolysis and exposing miR-7 to cytoplasmic nucleases (Shi et al. 2020). The degradation of miR-7 led to the derepression of α-synuclein and inflammation. Investigating the interplay of different classes of ncRNAs in controlling downstream targets and thus pathologic events like inflammation leads to identifying potential mechanisms to design new therapeutic targets to improve post-stroke outcomes.

Epitranscriptomics modulates post-stroke inflammation

RNAs can undergo >170 post-transcriptional modifications that regulate the stability, localization, translation efficiency, and interactions with other molecules of RNAs (Wiener & Schwartz 2021). These chemical modifications of RNA are known as epitranscriptomic changes, analogous to epigenetic modifications in DNA and histones (Nachtergaele & He 2018). Epitranscriptomic modifications are regulated by a diverse set of enzymes known as writers, readers, and erases. These mediate the addition, recognition, and removal of the specific chemical moieties on RNA, respectively (Gilbert et al. 2016). N6-methyladenosine (m⁶A), N1-methyladenosine (m¹A), 5-hydroxymethylcytosine (5hmc), 5-methylcytosine (5mc), pseudouridine (ψ), and inosine are the major epitranscriptomic modifications in mammals (Song & Yi 2017). Epitranscriptomic modifications are essential for normal development, and their dysregulation underlies the etiopathology of several diseases, including stroke (Chokkalla et al. 2022a).

The m⁶A is the most prevalent and reversible mRNA modification in the adult mammalian brain (Meyer et al. 2012) that regulates the structure, metabolism, localization, translation, and decay of mRNAs (Livneh et al. 2020). Several RNAs were observed to be m⁶A modified in the peri-infarct cortex of mice subjected to transient focal ischemia (Chokkalla et al. 2019). These include both mRNAs (122 hyper- and 5 hypo-methylated) and lncRNAs (17 hyper- and 3 hypo-methylated) (Chokkalla et al. 2019). Many pro-inflammatory mRNAs, including TNF-α, C-C motif chemokine ligands 2,3,4, and 6, IL-6, E-selectin, and MMP-3, were significantly hypermethylated indicating a role for m⁶A in post-stroke inflammatory response (Chokkalla et al. 2019). Increased m⁶A methylation has a functional significance as overexpression of m⁶A demethylase, fat Mass and obesity-associated protein (FTO) led to suppression of mRNA hypermethylation and decreased post-stroke infarct volume, motor deficits, cognitive decline, and depression-like behavior in mice independent of sex (Chokkalla et al. 2023). A recent study also showed that FTO overexpression counteracted microglial response after transient cerebral ischemia in mice by attenuating cGAS-STING signaling, leading to reduced expression of pro-inflammatory mediators TNF-α and IFN-β (Yu et al. 2023).

The m¹A is another prevalent epitranscriptomic modification, which is an isomer of m⁶A with methylation at the N1 instead of the N6 position (Jin et al. 2022). The post-transcriptional gene regulation by m¹A is site and context-specific. For instance, the m1A mark in coding sequence regions of mRNAs can inhibit translation, whereas m¹A in the 5’UTR promotes translation (Li et al. 2017; Safra et al. 2017). The m¹A enrichment is highest in the brain compared to other organs, and altered m¹A tagging to RNAs has been implicated in the etiology of Alzheimer’s disease (Shafik et al. 2022; Lee et al. 2007).

In the context of ischemic stroke, we observed a differential m¹A tagging on mRNAs (8 transcripts hypermethylated and 6 were hypomethylated) in the peri-infarct cortex of mice following transient cerebral ischemia (Chokkalla et al. 2022b). Interestingly, most m1A modifications are found in the 3’-UTR of the unannotated transcripts transcribed from the vicinity of the genes regulating protein complex assembly, circadian rhythms, chromatin remodeling, and chromosome organization (Chokkalla et al. 2022b). A recent study reported that oxygen-glucose deprivation/reperfusion injury differentially induces m¹A marks on mRNAs, lncRNAs, and circRNAs in primary neurons (Zhang et al. 2023). This altered m1A landscape in primary neurons is shown to be associated with the pathways regulating mast cell chemotaxis, autophagy, and nucleic acid catabolism, suggesting the implication of m¹A modification in neuronal homeostasis under hypoxic stress (Zhang et al. 2023).

While the functional significance of m¹A in ischemic brain injury is yet to be explored, differential expression of m¹A regulatory genes (upregulated: TRMT61A, RRP8, YTHDC-1, YTHDF-1, 2 and 3 & downregulated: FTO and ALKBH1) positively correlated to increased inflammatory cell infiltration in human abdominal aortic aneurysm tissues, suggesting a link between m¹A dysregulation and inflammation (Wu et al. 2022). Other less abundant RNA modifications like 2′-O-methylation (Nm) and N7-methylguanosine (m7G) also decreased in the post-stroke mouse brain, but the functional significance of these epitranscriptomic changes in pathophysiologic events such as inflammation warrants further studies (Chokkalla et al. 2022b).

Role of epigenetics in post-stroke neuroinflammation

Epigenetics involves the interaction of external stimuli with DNA and histones to modify gene expression without changing the DNA sequence. Disruption of several epigenetic mechanisms, including DNA methylation and hydroxymethylation, histone modifications, and the interaction with ncRNAs, influence post-stroke pathophysiologic mechanisms, including inflammation, and thus alters the susceptibility and recovery after stroke (Morris-Blanco et al. 2019; Morris-Blanco et al. 2021; Phillips et al. 2023; Morris-Blanco et al. 2022a).

Cytosines in the DNA are methylated by DNA methyltransferases (DNMTs) DNMT1, DNMT3a, and DNMT3b, which silences the gene expression. This process predominantly occurs in the CpG islands (>500 base pairs with >55% CG content) in the promoters throughout the genome (Takai & Jones 2002). Many factors, such as sex, age, and folate status, and physiologic insults, such as excessive inflammation and oxidative stress, can potentially impact DNA methylation (Kim et al. 2009).

The overall DNA methylation levels, although, are increased in ischemic brain tissue and may be responsible for promoting cell death following an ischemic stroke in mice, several clinical and preclinical studies on stroke have identified aberrant DNA methylation patterns suggesting that hypo- or hypermethylation at specific gene-promoter sites may have differential effects on stroke outcomes (Endres et al. 2000; Gallego-Fabrega et al. 2016; Baccarelli et al. 2010; Asada et al. 2020). For example, patients that are at an increased risk for stroke have hypomethylation of Long Interspersed Nucleotide Element-1 (LINE-1) repeats associated with the elevated levels of serum VCAM-1, lower TRAF3 DNA methylation levels correlated with increased platelet aggregation and vascular recurrence of ischemic stroke and increased DNMT3a activity as a predictor of stroke outcome (Baccarelli et al. 2010; Gallego-Fabrega et al. 2016; Asada et al. 2020). On the other hand, hypermethylation of the gene encoding S-adenosylhomocysteine hydrolase (AHCY) was observed in both male and female ischemic stroke patients than in control groups (Zhao et al. 2020).

In rodents, cerebral ischemia induces DNA methylation that is thought to promote cell death (Endres et al. 2000). In support, DNMT inhibition was shown to protect the post-stroke brain (Pandi et al. 2013; Endres et al. 2000; Endres et al. 2001). Following the stroke, many differentially methylated genomic regions were observed in cerebral microvessels, including genes that code for structural proteins, transporters and channels, and proteins involved in endothelial cell processes (Phillips et al. 2023). In the post-stroke brain, hypermethylation of the Timp-2 promoter prevented Timp-2 expression, leading to increased MMP-9 activity and permeability of the blood-brain barrier, which can be reversed by inhibiting DNMT (Phillips et al. 2023; Endres et al. 2000; Choi et al. 2022). These studies highlight the role of DNA methylation in regulating inflammation and BBB function following stroke (Phillips et al. 2023).

The 5-methylcytosine (5mC) will be converted to 5-hydroxymethylcytosine (5-hmC) by ten-eleven translocation dioxygenases (TET1, TET2 and TET3). The 5hmC promotes gene expression. We recently reported that focal cerebral ischemia in adult rodents increases DNA hydroxymethylation of several neuroprotective genes in a TET3-dependent manner in the peri-infarct cortex, and TET3 knockdown exacerbates secondary brain damage in both sexes (Morris-Blanco et al. 2021; Morris-Blanco et al. 2019). Furthermore, induction of TET3 by ascorbate increases 5hmC in the promoters of anti-inflammatory genes and improves functional recovery in mice independent of sex (Morris-Blanco et al. 2022b). TET2-mediated increase in 5hmC was also shown to attenuate the expression of pro-inflammatory genes ICAM-1, VCAM-1, IL-1β and MCP-1 in atherosclerotic plaques of high-fat diet-fed ApoE knockout mice (Peng et al. 2016). Furthermore, TET2 and/or TET3 knockdown led to increased ischemic brain injury (Morris-Blanco et al. 2021; Morris-Blanco et al. 2019; Miao et al. 2015).

Epigenetic control of inflammation following stroke also involves alterations in gene expression through post-translational modifications of histones that include acetylation, methylation, phosphorylation, and ubiquitination. Among these, histone acetylation regulated by histone acetyltransferases and HDACs is extensively studied. HDAC inhibitors such as trichostatin A, valproic acid, sodium butyrate, sodium 4-phenylbutyrate, and suberoylanilide hydroxamic acid have been shown to provide robust protection against post-stroke inflammation by curtailing the expression of inflammatory molecules such as IL-1β, IL-17A and IL-18 (Park & Sohrabji 2016; Langley et al. 2009; Lu et al. 2023).

All the above studies suggest that epigenetic control of post-stroke inflammation is a dynamic and intricate process that influences stroke outcomes. Further understanding of the epigenetic mechanisms provides new targets to develop therapies to minimize brain damage and promote functional recovery after stroke.

Perspective, therapeutics and conclusions

Post-stroke neuroinflammation is crucial for clearing debris to prepare the brain for plasticity and repair. However, prolonged inflammation exacerbates secondary brain damage and impairs tissue repair. Following stroke, activated endothelial cells release specific molecules such as selectins and adhesion molecules, encouraging the attachment and movement of immune cells across the blood-brain barrier. The infiltrated immune cells in the ischemic brain release various molecules, such as cytokines/chemokines, complement proteins, and MMPs, that promote inflammation. Additionally, the immune response within the ischemic parenchyma is initiated by the release of DAMPs that immune system effectors like microglia detect through their pattern recognition receptors.

This cascade of events activates multiple transcription factors, including NF-κB, AP-1, IRFs and STATs, in different brain cell types. These transcription factors thus differentially modulate post-stroke gene expression, resulting in inflammation, neuronal damage/survival, and outcomes. However, the functioning of these factors in the post-stroke brain also depends on the crosstalk between them, accessibility to DNA binding sites, action of ncRNAs, and interplay of other epigenetic factors and modifications such as histone modifications to control gene expression responsible for post-stroke inflammation.

Various ncRNAs, including miRNAs, lncRNAs, and circRNAs, have emerged as essential regulators of gene expression in different biological processes, including post-stroke inflammation. Moreover, post-stroke modifications to RNA molecules (epitranscriptomics) impact gene expression and protein synthesis, influencing inflammation. Particularly, m6A, the most common type of RNA modification, modulates various molecular processes such as RNA stability, translation, microRNA biogenesis, and splicing, impacting inflammation, for instance, by polarizing the microglial state. Thus, manipulating epitranscriptomic marks may offer potential avenues to regulate the inflammatory response, promote tissue repair, and enhance functional recovery after stroke. Additionally, epigenetic modifications that impact inflammation are crucial for post-stroke etiology and recovery. Mechanisms such as differential DNA methylation, hydroxymethylation, and/or histone modification further play a decisive role in regulating gene expression essential for modulating inflammatory response and post-stroke outcome.

Interestingly, pharmacomodulation and/or genetic manipulation of transcription factors, ncRNAs, and epitranscriptomic and epigenetic regulators are being actively pursued to develop novel therapeutics. These include activators/inhibitors of transcription factors, epigenetic modulators such as DNMTs, TETs, and HDACs and RNA-based therapeutics. We recently observed that ascorbate induces TET3 activity thereby increasing 5hmC levels in the promoters of several anti-inflammatory genes, suggesting the potential of epigenetic reprogramming as an innovative approach to mitigate post-stroke injury (Morris-Blanco et al. 2022b). Over the past decade, a significant effort has been made toward developing RNA therapies, and many of those received FDA approval for various CNS diseases (Anthony 2022; Mendell et al. 2017). Several approaches involving the manipulation of ncRNAs such as miRNA miR-7 and lncRNAs FosDT and Malat1 were shown to improve the post-stroke outcomes in animal models (Kim et al. 2018; Mehta et al. 2023a; Mehta et al. 2023c; Wang et al. 2020a). RNA therapeutics are in their early phase of development with many challenges including specificity, delivery, and tolerability. In addition, other factors such as the role of RNA modifications, RNA processing, RNA-RNA, and RNA protein interaction in modulating post-stroke inflammatory pathophysiology need more research to develop novel therapeutic strategies (Winkle et al. 2021). For instance, m⁶A methylation or RNA can be altered by manipulating FTO activity with non-steroidal anti-inflammatory agent meclofenamic acid (MA) or NADPH (Wang et al. 2020b; Huang et al. 2015). It is particularly important since the decrease in m⁶A methylation and post-stroke recovery can be facilitated by FTO overexpression (Chokkalla et al. 2023). Both epigenetic modifications of DNA and epitranscriptomic modifications of RNA respond to conditions like stroke and thus influence the functional outcome. Hence, the precise curation of their function will play a crucial role in therapeutic development.

Succinctly, the interplay among transcription factors, ncRNAs, and epitranscriptomic and epigenetic changes orchestrate post-stroke inflammation, thereby regulating secondary ischemic brain damage. Gaining a deeper insight into these mechanisms is thus crucial in identifying novel targets and developing therapeutic strategies to mitigate post-stroke neurologic dysfunction and facilitate functional recovery.

List of abbreviations

3’-UTR

3’-untranslated region

5hmc

5-hydroxymethylcytosine

5-hmC

5-hydroxymethylcytosine

5mc

5-methylcytosine

5mC

5-methylcytosine

AHCY

S-adenosylhomocysteine hydrolase

ATF-2

activating transcription factor-2

ATP

adenosine triphosphate

BBB

blood-brain barrier

Bcl-2

B-cell lymphoma 2

BMECs

brain meningeal endothelial cells

BMSCs

bone marrow mesenchymal stem cells

C/EBP

CCAAT enhancer binding protein

C1QTNF6

complement C1q tumor necrosis factor-related protein 6

circRNAs

circular RNAs

CREB

cAMP response element-binding protein

DAMPS

damage-associated molecular patterns

DNA

deoxyribonucleic acid

DNMTs

DNA methyltransferases

Egr1

early growth response-1

ER

endoplasmic reticulum

eRNAs

enhancer RNAs

FosDT

Fos downstream transcript

FTO

fat mass and obesity-associated protein

Gli2

GLI family zinc finger 2

GM-CSF

granulocyte-macrophage colony-stimulating factor

HDAC1

histone deacetylase 1

HIF-1

hypoxia-inducible factor-1

HMGB1

high mobility group box 1

HSPs

heat shock proteins

ICAM-1

intercellular adhesion molecule 1

IFN-γ

interferon-γ

ILs

interleukins

iNOS

inducible nitric oxide synthase

lincRNAs

large intergenic ncRNAs

LINE-1

long interspersed nucleotide element-1

lncRNAs

long noncoding RNAs

m1A

N1-methyladenosine

m6A

N6-methyladenosine

m7G

N7-methylguanosine

MCP-1

monocyte chemoattractant protein-1

MIP-1α

macrophage inflammatory protein-1α

miRNAs

microRNAs

MMPs

matrix metalloproteases

MREs

miRNA response elements

MyD88

myeloid differentiation primary response-88

ncRNAs

noncoding RNAs

NF-κB

nuclear factor kappa B

NIHSS

NIH Stroke Scale

NLRP3

NOD-like receptor pyrin domain containing 3

Nm

2′-O-methylation

piRNAs

PIWI-interacting RNAs

Pol II

RNA polymerase II

PPAR

peroxisome proliferator-activated receptors

pri-miRNA

primary miRNA

RBPS

RNA binding proteins

RISC

RNA-induced silencing complex

RNA

ribonucleic acid

ROS

reactive oxygen species

rRNAs

ribosomal RNAs

snoRNAs

small nucleolar RNAs

snRNAs

small nuclear RNAs

Spi1

spleen focus forming virus proviral integration oncogene

STAT

signal transducer and activator of transcription

TETs

ten-eleven translocation dioxygenases

Tlr4

toll-like receptor 4

TNF-α

tumor necrosis factor- α

tPA

tissue plasminogen activator

TRAF

TNF-receptor associate factor

tRNAs

transfer RNAs

VCAM-1

vascular cell adhesion molecule 1

YAP

Yes-associated protein

Ψ

pseudouridine