Post-stroke inflammation is an integral contributor to secondary injury progression, and an excessive inflammatory response could escalate the second wave of neural cell death and deteriorate the outcome of stroke [38]. Controlling the destructive inflammatory response is a feasible strategy for stroke therapy. The present study demonstrated that EDB exerted powerful anti-inflammatory activity in an experimental mouse model of stroke, which was linked to its regulation of central microglia and astrocyte and the influx of peripheral immune cells.
Although the neuroprotective effect of EDB on acute cerebral ischemia has been previously evidenced in a rat model [12], we still performed TTC staining and MRI to verify the effectiveness of EDB against ischemia in an experimental mouse model. We analyzed brain infarction in each 1 mm coronal section from Bregma 2.1 mm to Bregma 3.08 mm by TTC staining and found that the alleviation of the infarct region in the EDB groups mainly involved the primary motor cortex (M1), caudate putamen (CPu, striatum), primary somatosensory cortex (S1), primary visual cortex (V1) and part of the hippocampus region by comparing the maps of the mouse brain. These findings indicated that EDB treatment could ameliorate stroke-induced neurologic deficits in motor function, sensation, vision, and cognition, which are controlled by specialized regions in the brain. Corresponding to the TTC results, EDB treatment improved motor coordination and somatosensory recovery, as determined by a battery of behavioral assessments. In addition, the temporal development of infarction growth was estimated through MRI imaging and, together with behavioral tests, indicated the long-term effects of EDB on functional recovery.
Next, we investigated the inflammatory changes during the early phase after MCAO in the presence or absence of EDB. In accordance with the results in qRT-PCR, proinflammatory cytokines or chemokines, including IL-6, IL-1β, TNF-α, IFN-γ, CXCL1, and MCP-1, were quickly induced by MCAO insult and remained high for 72 h, which is consistent with a previous study [39]. Systemically administered EDB attenuated the expression of these proinflammatory mediators in the ischemic brain, suggesting that EDB-afforded neuroprotection partially relied on its anti-inflammatory activity. However, EDB failed to up-regulate the levels of anti-inflammatory cytokines, including TGF-β and IL10, which are sub-acutely elevated with time following stroke onset. These cytokines or chemokines have been identified to be implicated in infarct development and have been proposed as potential targets or biomarkers for stroke management [40, 41]. IL-1β is a prototypical inflammatory cytokine that may exacerbate brain damage by promoting neuronal apoptosis, potentiating persistent inflammation, and inducing the expression of adhesion molecules [42]. Unlike IL-1β, which exerts neurotoxic effects, TNF-α has been considered a neurotoxic or neuroprotective player in the progression of ischemia [43, 44]. IL-6 is regarded as a pleiotropic cytokine that plays an ambivalent role in ischemic brain pathogenesis, acting as an inflammatory mediator in the acute period and working as a neurotrophic factor during the subacute and later stages [45, 46]. TGF-β is a growth factor with anti-inflammatory properties that blocks pro-inflammatory cytokine production [47, 48], and the cytokine IL-10 is known as a key mediator with anti-inflammatory and anti-apoptotic effects. MCP-1 is a chemokine that is involved in monocytic recruitment, and CXCL1 was reported to be associated with the migration of neutrophils [49–51]. The effect of IFN-γ on neuroinflammation is intricate, it acts on multiple cell types and may depend on the specific microenvironment [52], and it seems to play a pivotal role in maintaining the integrity of the BBB [53]. These cytokines or chemokines are generated by various immune or non-immune cells, and the major cellular sources are based on the timing and context.
Microglia are the primary immune cells in the regulation of neuroinflammation and exhibit a pro- (M1) or anti-inflammatory (M2) state during all phases of stroke. Therefore, we investigated the phenotypic alterations in microglia in the presence of EDB. In accordance with the results in flow cytometry and immunofluorescence, EDB inhibited the activation of microglia towards the M1 phenotype, as indicated by CD68, CD86, and CD16/32 expression, but did not promote microglial polarization to the M2 phenotype, as indicated by CD206 expression. Similar changes in CD68 and CD206 mRNA levels were observed in sorted microglia by qRT-PCR. Correspondingly, EDB application inhibited the generation of M1 cytokines (IL-1β, TNF-α,) but did not affect M2 cytokines (TGF-β and IGF-1), even though there were slightly elevated mRNA levels of IGF-1. Notably, in contrast to the elevated expression of IL-1β and TNF-α after MCAO, IL-6 mRNA levels in isolated microglia were sharply decreased 3 d after MCAO, as previously indicated [54]. The possible reason for this phenomenon may be that the time point selected was too late, and an ischemia-induced increase in IL-6 occurred from 6 h to 12 h poststroke [45, 55].
In addition to microglia, astrocytes are vital participants in regulating neuroinflammation and exert pro-inflammatory or anti-inflammatory effects by the A1 or A2 phenotype. Lcn-2 is a vital marker of reactive astrocytes that is involved in promoting pro-inflammatory activation of astrocytes [56]. Lcn-2 gene expression can be highly upregulated in both MCAO- and LPS-stimulated reactive astrocytes [30], and Lcn-2 deficiency provides protective effects against brain injury by limiting the classical proinflammatory phenotypes of activated astrocytes and dampening the production of inflammatory mediators [57–59]. Serping1 is a crucial regulator of the complement cascade [60] and is prominently expressed in LPS-induced reactive astrocytes (A1 phenotype). S100a10 is commonly recognized as an A2 phenotypic gene that is greatly induced in MCAO-induced reactive astrocytes but not LPS-induced reactive astrocytes. In the current study, ischemia-induced gene expression of Lcn-2 and Serping1 in astrocytes was markedly downregulated by EDB, which indicates that EDB can restrain the activation of A1-type astrocytes. Furthermore, we showed that EDB significantly decreased the mRNA levels of the proinflammatory cytokines IL-6, IL-1β, and TNF-α, which were upregulated in A1 astrocytes. However, EDB therapy did not change the polarization of A2-type astrocytes, as indicated by the phenotypic gene S100a10 and anti-inflammatory gene TGF-β.
Collectively, we found that EDB alleviated neuroinflammation by modulating the polarization of microglia and astrocytes in the MCAO mouse model. However, factors that could affect the phenotypic polarization of glial cells are multifaceted under the pathological conditions of ischemia. For example, the interaction between microglia and astrocytes plays a role in phenotypic change. On the one hand, microglia-derived IL-1β, TNF-α, and IL-6 can induce A1 reactive astrocytes, which amplify the cascade of neuroinflammation [28]. On the other hand, M2 microglia-produced IL-10 may promote astrocyte switching to the A2 phenotype, which in turn upregulates TGF-β expression and reduces the activation of microglia [61]. The direct anti-inflammatory effect of EDB on microglia and macrophages has been reported in BV2 [62] and RAW264.7 cells [63]. However, whether the regulatory effect of EDB on astrocytes is direct or indirect has not been fully explored and requires further research. As a previous study demonstrated that NO-induced apoptosis in primary astrocytes could be inhibited by edaravone [64], we speculate that EDB may have a direct effect on astrocytes.
After the activation of brain-resident glial cells following ischemia, peripheral leukocytes are recruited into the ischemic brain and contribute to exacerbate brain tissue damage [65]. The reduced production of CXCL1 and MCP-1 in the cortex hints that EDB may decrease the influx of inflammatory cells. Actually, EDB significantly diminished the invasion of peripheral immune cells (e.g., macrophages, NK cells, neutrophils, and monocytes). However, systemic EDB administration did not alter the composition of macrophages, monocytes neutrophils, NK cells, and T and B lymphocytes in the blood and spleen. Considering that borneol has been reported to play a role in maintaining the integrity of the BBB [66], we investigated the influence of EDB on tight junction expression and found that EDB alleviated MCAO-induced deterioration of the tight junction proteins VE-cadherin, ZO-1, claudin-5 and occludin at the mRNA level, which suggests that the decrease in leukocyte infiltration induced by EDB is potentially related to its protective action on the BBB. Among brain-recruited immune cells, macrophages play a predominant role in modulating the neuroinflammatory response, and the quantity of macrophages is greater than that of other cells. Further analysis showed that EDB significantly diminished the accumulation of CD86+(M1) macrophages in the ischemic brain, which is largely associated with the restrained polarization of M1 macrophages in the spleen and blood. The mRNA expression of IL-1β, TNF-α, and IL-6 in infiltrated macrophages was also hindered by EDB, which demonstrates that EDB could modulate the neuroinflammatory events orchestrated by recruited macrophages. Similar alterations were observed in peripheral macrophages from the spleen and blood, suggesting that the polarization of M1 macrophages was probably suppressed before the cells migrated to brain tissue. Regretfully, the function of brain-infiltrated monocytes, neutrophils, and NK cells in the inflammatory events after EDB administration was not elucidated in this study, and needs further research.