Elevated microRNA-129-5p level ameliorates neuroinflammation and blood-spinal cord barrier damage after ischemia-reperfusion by inhibiting HMGB1 and the TLR3-cytokine pathway

Spinal cord ischemia-reperfusion (IR) injury is a severe but unpredictable and unpreventable complication that commonly occurs during thoracoabdominal aortic surgery. IR-induced inflammatory responses, including increases in extracellular matrix degradation and vascular permeability, can aggravate spinal cord edema and worsen neurologic deficits, resulting in a high incidence of paraplegia [1, 2]. Treatments that might regulate the above mechanisms have become a focus in the clinic [3, 4]. Previous studies have shown that increases in microglia and upregulation of Toll-like receptors (TLRs) are associated with proinflammatory cytokine release through recognition of damage-associated molecular pattern molecules (DAMPs) during spinal cord ischemia [4, 5]. High-mobility group box-1 (HMGB1) is a well-characterized DAMP that is conserved in most mammals [6]. Under physiological conditions, HMGB1 functions to stabilize nucleosomes. In various models of injury, HMGB1 is immediately released by injured cells to establish and amplify inflammatory responses via receptor for advanced glycation end products (RAGE) or TLRs through the release of chemokines and cytokines [7‐10]. We have demonstrated that activation of the TLR4-mediated nuclear factor-kappa B (NF-κB)/interleukin (IL)-1β positive feedback loop promotes substantial inflammatory damage to neurons and the blood-spinal cord barrier (BSCB) after spinal cord IR [5]. Moreover, the hyper-secreted HMGB1 is detected before the secretion of tumor necrosis factor (TNF)-α and IL-1β during spinal cord injury [7]. Therefore, we hypothesize that HMGB1 is an upstream modulator of the TLR-mediated pathway in the spinal cord. Treatments that regulate the interactions between TLRs and HMGB1 may be used as novel therapeutic approaches for spinal cord IR injury.

MicroRNAs (miRs) are 20–22-nucleotide-long non-coding RNAs that are specifically expressed in certain organs or cells to negatively regulate target gene expression at the posttranscriptional level [11, 12]. Many miRs have been shown to be dysregulated during central nervous system (CNS) injury [11, 12]. Recently, miR-129-5p, which is known to function as a tumor suppressor [13], has been shown to function in neuro-apoptosis by regulating bcl-2 and caspase-3 expression during cerebral IR [11]. miR-129-5p has also been shown to modulate angiogenesis and revascularization by regulating vascular endothelial growth factor expression via RAGE-HMGB1 signaling during intracerebral hemorrhage [14]. Additionally, miR-129-5p prevents NF-κB transduction and subsequent inflammatory infiltration in autoimmune diseases by inhibiting TLR2 or TLR4-HMGB1 signaling [10, 15]. TLR3, which was originally identified as the receptor for viral RNA, has been demonstrated to be essential for HMGB1 in facilitating agonists with other TLRs during neuroinflammation and neurodegeneration [16, 17]. Upregulation of TLR3 stimulates NF-κB transcription and increases the expression of neurodegeneration markers [16]. In contrast, TLR3-deficient mice are resistant to galactosamine sensitization and display reduced production of inflammatory cytokines [18]. However, little is known about the function of miR-129-5p in the interaction between HMGB1 and TLR3 in mouse models of spinal cord IR. In this study, first, we investigated whether miR-129-5p was dysregulated and then defined HMGB1 as its target in this context. Next, we assessed the effects of miR-129-5p on HMGB1 expression and the TLR3-cytokine pathway in vivo by intrathecal injection of an miR mimic and control, recombinant HMGB1 (rHMGB1), and a TLR3-specific agonist (polyinosinic-polycytidylic acid, Poly(I:C)). Additionally, the effects of these treatments on hind-limb motor function and BSCB leakage were evaluated to determine the role of miR-129-5p in spinal cord IR injury.

Various miRs that are widely distributed in the CNS have been shown to play essential roles in modulating the pathogenesis of IR injury, thereby contributing to disease development and outcomes [10, 12]. In this study, we provided evidence for the first time that elevated miR-129-5p expression exerted neuroprotective effects in a mouse model of spinal cord IR.

miR-129-5p is known to regulate malignant tumor progression and metastasis [13]. Recently, emerging evidence has shown that miR-129-5p is dysregulated during trauma and degenerative and autoimmune diseases and thus has special significance for the maintenance of neuronal function and macrophage/monocyte migration under these pathological conditions [10, 14, 22]. Transfection of human osteoblast-like cells with an miR-129-3p mimic inhibits in vitro monocyte migration by targeting the human IL-17 gene, which is one of the major mechanisms underlying the pathogenesis of rheumatoid arthritis [22]. A study by Sepramaniam et al. suggested that among the various biomarkers identified, miR-129-5p was a significantly altered miR in stroke patients [23]. These findings indicate that miR-129-5p may be involved in regulating neuroinflammatory responses after injury. In this study, we demonstrated that miR-129-5p was downregulated in a temporal fashion and reached its lowest levels at both 12 and 48 h (Fig. 2), which was consistent with the results of our previous studies demonstrating the bimodal stages of IR-induced neuroinflammation and suggesting potential miRNA-based treatment strategies [21, 24].

We previously showed that some miRs are associated with TLR signaling pathways in IR [2]. Each miR can affect multiple target mRNAs, and each target mRNA can also be regulated by several miRs [1, 2, 20]. We used TargetScan to identify the targets of miR-129-5p by specifying a continuous “seed match” of more than six base pairs and a potential link with TLR signaling (e.g., adaptor, interacting proteins, and DAMPs). In a rat model of intracerebral hemorrhage, miR-129-5p has been shown to be associated with the inhibition of revascularization and angiogenesis by suppressing HMGB1 and RAGE signaling [14]. Thus, HMGB1 was selected as a potential target of miR-129-5p. The temporal expression patterns of HMGB1 and miR-129-5p preliminarily indicated a negative correlation between these two factors. This interaction was confirmed using a luciferase assay; a reduction in luciferase activity was observed in cells co-transfected with a vector containing the WT 3′UTR of HMGB1 and mimic-129 (Fig. 3). Furthermore, to account for the in vivo complexity and crosstalk, the above effects were also evaluated in a mouse model by pretreating the mice with mimic-129 and con-129 [2, 22]. Intrathecal injection of mimic-129 markedly reduced HMGB1 mRNA and protein expression, whereas these changes were not observed after treatment with con-129 (Fig. 5). Collectively, these data suggest that miR-129-5p functions as a negative modulator of HMGB1 in spinal cord IR.

HMGB1 is a highly conserved DNA-binding protein that functions as an intracellular regulator of nucleosome formation and gene transcription [25]. Recently, extracellular HMGB1 has been shown to be a potent proinflammatory factor that can lead to cytokine release and tissue damage [7, 9, 25]. Previous studies have demonstrated that HMGB1-mediated inflammation is a major mechanism in IR-induced brain damage [26]. Motor neurons are the most vulnerable to IR, and thus IR injury is usually associated with motor function impairments [21]. Using electron microscopy, Chen et al. demonstrated that translocation of HMGB1 from the nucleus to the cytoplasm in spinal neurons occurred within a few hours after trauma [7]. Additionally, the BSCB is critical for maintaining spinal homeostasis [4, 5]. Therefore, the role of HMGB1 in inflammatory responses against neurons and BSCB is most aptly represented by its specific cellular distribution. Spinal microglia, which are a type of innate immune cell, function as macrophages and are capable of phagocytosis immediately after IR [4]. Consistent with the study by Chen et al. [7], the immunofluorescence images in Fig. 4 show that overexpressed HMGB1 is mainly distributed in the cytoplasm of neurons and microglia. Additionally, the effects of miR-129-5p were measured at 48 h post-surgery when the maximal immunoreactivity against HMGB1 was observed. Compared with the IR group, the mimic-129 group had a markedly reduced number of double-labeled neurons and microglia, but these cells were relatively unchanged in the con-129 group, confirming that HMGB1 was released from stimulated neurons and activated microglia after IR (Fig. 6). Consistently, obvious improvement in hind-limb motor function and preservation of BSCB integrity were observed in the mimic-129 group, as demonstrated by elevated Basso Mouse Scale scores and decreased EB extravasation and water content (Fig. 7).

TLR3 is a unique member of the TLR family that is located in endosomal compartments or on the cell surface [27]. TLR3 usually recognizes pathogen-derived nucleic acids and recruits an adaptor protein to induce interferon-ß release [17, 28]. Recently, several studies investigated the essential roles of TLR3 activation in the amplification of inflammatory responses and cytokine release in response to HMGB1 [17, 29, 30]. Consistently, the increased immunoreactivity of HMGB1 in the cytoplasm coincided with the fluorescence signal representing TLR3 on the cell surface (Fig. 5). These increases in the translocation of TLR3 to the cell surface expose TLR3 to stimuli produced by injured cells [31]. In this study, intrathecal injection of rHMGB1 and Poly(I:C) both induced comparably higher levels of HMGB1 and TLR3 expression and led to robust secretion of IL-1β and TNF-α and BSCB leakage compared to the levels detected in the untreated IR group. These results are consistent with those of an in vivo study on alcohol-sensitized neuroinflammation [16]. In that study, sequential treatment with ethanol and Poly(I:C) facilitated the effects of Poly(I:C) on the induction of TLR3 and HMGB1 expression [16]. In another study, delayed release of IL-1β and TNF-α was observed in monocytes cultured with HMGB1 compared with monocytes directly treated with the TLR4 agonist lipopolysaccharide [7]. These findings indicate that HMGB1 may be a trigger for TLR3 activation. Early amplification of inflammatory responses can directly activate downstream signals without requiring HMGB1. Nevertheless, inhibition of HMGB1 with mimic-129 downregulated the expression of TLR3 and downstream cytokines, confirming that HMGB1 was required for TLR3 activation. In contrast, in a study by Yamagata et al., HMGB1 was not upregulated in astrocytes in response to stroke [32]. This difference might be attributable to the lack of effect of HMGB1 on intracellular NF-κB localization in cultured astrocytes because HMGB1 was shown to decrease the translocation of NF-κB to the nucleus during brain edema formation [33]. In addition to HMGB1, some indirect factors may also be involved in the IR injury process. Given the complexity and crosstalk among glial cells and neurons, further in vitro BSCB models need to be constructed to elucidate the underlying mechanisms and uncover new treatment strategies.

In conclusion, this study explored the roles of miR-129-5p and its targets in spinal cord IR. Intrathecal pretreatment with mimic-129 protected neurons and the BSCB against inflammatory responses by downregulating HMGB1 expression and inhibiting TLR3-cytokine pathway activation. These results suggest the potential use of miRNA-based strategies for IR.