Minocycline increases the number of microglia and decreases microglial damage in the ischaemic penumbra
We then performed Iba-1 immunohistochemistry (IHC) and immunofluorescence (IF) staining to evaluate the number and morphology of microglia in the ischaemic penumbra (IP). The results showed that CIRI caused microglial activation and proliferation in the IP of MCAO/R rats. Microglia were enlarged and amoeboid after MCAO/R induction. Interestingly, the number of microglia in the minocycline-treated MCAO/R group was significantly greater than that in the MCAO/R groups (Fig. 2A-B).
To further investigate the ultrastructural changes in microglia in the IP, we used TEM. In the sham group, the structure of mitochondria in the cytoplasm of microglia was generally normal (showed prominent cristae and an intact membrane). CIRI caused remarkable microglial injury in the MCAO/R group, including the rupture of the outer mitochondrial membrane, the reduction or disappearance of mitochondrial cristae and vacuolation, without obvious alterations in the nucleus. Minocycline significantly ameliorated these morphological changes (Fig. 2C). Altogether, these data suggested that minocycline ameliorated microglial injury and promoted microglial numbers in IP.
Minocycline alleviates microglial death in vitro
To determine the reason for the increase in microglial numbers caused by minocycline treatment, an in vitro model of OGD/R was established in BV-2 cells to mimic microglial injury in CIRI. The results showed that minocycline could not induce microglial proliferation in the control group. Compared to the OGD/R group, cell viability was increased significantly in the minocycline treatment group. To further clarify whether cell death was rescued by minocycline, lactate dehydrogenase (LDH) release was tested. The results showed that LDH level in the minocycline treatment group was significantly lower than that in the OGD/R group (Fig. 3A). Moreover, costaining of PI and calcein-AM indicated that minocycline reduced the number of dead cells (Fig. 3B). These results indicated that minocycline treatment ameliorated microglial death during CIRI in vitro.
Ferroptosis is a major form of the microglial death induced by OGD/R in vitro
The major types of programmed cell death (PCD) include apoptosis, autophagy, necroptosis and ferroptosis. Various pharmacological inhibitors of cell death, including apoptosis (Z-VAD-FMK, VAD), autophagy (3-methyladenine, 3-MA), necroptosis (necrostatin-1, Nec-1) and ferroptosis (ferrostatin-1, Fer-1), were used to identify the PCD type that occurs in OGD/R. The results indicated that, compared to the OGD/R group, Fer-1 treatment significantly improved cell viability and reduced LDH release (Fig. 4A). This result was consistent with costaining of PI and calcein-AM (Fig. 4B). These results indicated that ferroptosis might be the major form of microglial PCD in CIRI. To conform this, we measured intracellular labile Fe2+, which accumulates in ferroptotic cells. The results showed that the labile Fe2+ concentration was significantly higher in the OGD/R group than that in the control group, and could be rescued by Fer-1 treatment (Fig. 4C). Furthermore, lipid peroxidation was analysed by detecting C11-BODIPY 581/591 and by 4-HNE accumulation. ROS accumulation and lipid peroxidation increased in OGD/R, while Fer-1 significantly mitigated this increase (Fig. 4D-E). Glutathione peroxidase 4 (Gpx4), Fth and transferrin receptor (Tfr) are commonly used as indicators of ferroptosis, and our Western blot results showed that, compared with the control group, Gpx4 level was lower in OGD/R group, and Fer-1 increased its expression. Conversely, Fer-1 repressed the Fth accumulation induced by OGD/R, but not the accumulation of Tfr (Fig. 4F).
To further evaluate mitochondrial function, MitoTracker Red CMXROS staining and TEM were performed. Interestingly, the red fluorescence intensity of MitoTracker Red CMXROS in the OGD/R group declined greatly, and Fer-1 showed ability to reverse this decline (Fig. 4G). These data illustrated that Fer-1 significantly maintained the degree of mitochondrial membrane potential which was decreased by OGD/R treatment. Further morphological and microstructural details were obtained by using TEM, and Fer-1 alleviated the ruptures of the outer mitochondrial membrane and reduction or vanishing of mitochondrial cristae caused by OGD/R (Fig. 4H). In summary, these results indicate that ferroptosis is a major form of microglial death during OGD/R.
Minocycline inhibits ferroptosis of microglia with OGD/R in vitro
To analyse the effect of minocycline on ferroptosis of microglia, we added it into medium and mixed with a final concentration of 80 µM before routine culture or OGD/R treatment. The results showed that minocycline decreased the high labile Fe2+ level induced by OGD/R (Fig. 5A). The lipid peroxidation data indicated that minocycline decreased ROS accumulation and 4-HNE expression in microglia after OGD/R treatment (Fig. 5B-C). Similar to Fer-1, minocycline enhanced Gpx4 while suppressing Fth expression after OGD/R (Fig. 5D). Mito-Tracker Red CMXROS staining and TEM photos demonstrated that minocycline also contributed to the restoration of mitochondrial function, improving mitochondrial membrane potential and mitochondrial structures (Fig. 5E-F). Taken together, these data indicated that minocycline can effectively attenuate OGD/R-induced ferroptosis in microglia.
HO-1 is the potential target dictating to ferroptosis of microglia with OGD/R in vitro
To identify the mechanism by which minocycline ameliorates microglial ferroptosis, gene expression analyses were performed, and the results identified a total of 1509 differentially expressed genes between the control group and OGD/R group, with 645 genes downregulated, and 864 genes upregulated. In contrast, there were 456 downregulated and 144 upregulated genes between the OGD/R and minocycline treatment groups. KEGG analysis revealed an enrichment of genes involved in mineral absorption. The differential expression of the top 10 genes in our gene expression omnibus data was ranked. Intriguingly, the expression levels of haem oxygenase 1 (Hmox1) and Fth were significantly increased in the OGD/R group versus the control group, and the expression levels of both genes were significantly decreased in the minocycline treatment group (Fig. 6A). We then performed RT_qPCR for the Hmox1, Fth, Gpx4 and Tfr genes, and the results indicated that Hmox1 was repressed by minocycline (Fig. 6B).
Hemin, a HO-1 (encoded by the Hmox1 gene) inducer, was then pretreated with BV-2 cells at 20 µM for 24 h. The protective effect of minocycline was significantly attenuated in microglia after OGD/R exposure. In the hemin treated group, cell viability was lower and LDH release was higher than those in the OGD/R group (Fig. 6C). This result was consistent with costaining of PI and calcein-AM (Fig. 6D). The Western blot results revealed that HO-1 expression caused by OGD/R induction was increased following hemin treatment, and the minocycline response was not appreciable (Fig. 6E). The effect of minocycline on labile Fe2+ levels, ROS accumulation and 4-HNE expression was reversed by hemin (Fig. 6F-H). The Western blot results showed that minocycline increased Gpx4 and reduced Fth expression, and these effects were also stopped by hemin (Fig. 6I).
Furthermore, the protective role of minocycline on mitochondrial potential and mitochondrial ultrastructure was inhibited by the administration of hemin (Fig. 6J-K). Collectively, the protective effect of minocycline on OGD/R -induced ferroptosis was inhibited by the overexpression of HO-1.
HO-1 overexpression reduces the protective effect of minocycline in MCAO/R rats
To evaluate the role of HO-1 overexpression in the protective effect of minocycline in vivo, rats were pretreated for 3 days with hemin 50 mg/kg/d (Fig. 7A). In the area of IP, the increased expression of HO-1 protein was confirmed by Western blot (Fig. 7B). Strikingly, infarcted volumes were significantly higher in the hemin treatment group than in all the other groups (Fig. 7C-D), and all NDSs of the groups reached 3 (Fig. 7E). Moreover, the concentration of tissue iron was significantly increased after MCAO/R, particularly in the hemin treatment group (Fig. 7F).
HO-1 overexpression counteracts the protective effect of minocycline on microglial ferroptosis in vivo
Using IHC and IF staining of the IP, we found that the overall cellularity was significantly decreased in the hemin treatment group during MCAO/R. Cell numbers, including microglia, showed little variation even when minocycline was injected promptly for 72 h (Fig. 8A-B). Furthermore, upon preinjection of hemin, minocycline was able to partially decrease the overexpression of HO-1 induced by MCAO/R (Fig. 8C). The mitochondrial damage in microglia after MCAO/R was more severe than that in the other groups, and the protective effect of minocycline on mitochondrial structure was not as strong as that in the previous steps. Ruptures of the outer mitochondrial membrane, reductions or vanishing of mitochondrial cristae and vacuolation were still present in many microglia (Fig. 8D). Consequently, the protective effect of minocycline on MCAO/R was also inhibited by the overexpression of HO-1.