Minocycline Promotes Functional Recovery in Ischemic Stroke by Modulating Microglia Polarization Through STAT1/STAT6 Pathways

BACKGROUND
Increasing evidence suggests that microglia experience two distinct phenotypes after acute ischemic stroke (AIS): a deleterious M1 phenotype and a neuroprotective M2 phenotype. Promoting the phenotype shift of M1 microglia to M2 microglia is thought to improve functional recovery after AIS. Minocycline, a tetracycline antibiotic, can improve functional recovery after cerebral ischemia in pre-clinical and clinical research. However, the role and mechanisms of minocycline in microglia polarization is unclear.


METHODS
Using the transient middle cerebral artery occlusion - reperfusion (MCAO/R) model, we treated mice with saline or different minocycline concentration (10, 25, or 50 mg/kg, i.p., daily for 2 wk) at 24 h after reperfusion. Neurobehavioral evaluation, rotarod test, and corner turning test were carried out on day 14 after reperfusion. Then, neuronal injury, reactive gliosis, and microglia polarization were performed on day 7 following MCAO/R. Finally, we treated primary microglial cultures with LPS (Lipopolysaccharide; 100 ng/mL) plus IFN-γ (20 ng/mL) 24 h to induce M1 phenotype and observed the effects of minocycline on the M1/M2-related mRNAs and the STAT1/STAT6 pathway.


RESULTS
We found that a 14-day treatment with minocycline increased the survival rate and promoted functional outcomes evaluated with neurobehavioral evaluation, rotarod test, and corner turning test. Meanwhile, minocycline reduced the brain infarct volume, alleviated neuronal injury, and suppressed reactive gliosis on day 7 following MCAO/R. Moreover, we observed an additive effect of minocycline on microglia polarization to the M1 and M2 phenotypes in vivo and in vitro. In the primary microglia, we further found that minocycline prevented neurons from OGD/R-induced cell death in neuron-microglia co-cultures via regulating M1/M2 microglia polarization through the STAT1/STAT6 pathway.


CONCLUSION
Minocycline promoted microglial M2 polarization and inhibited M1 polarization, leading to neuronal survival and neurological functional recovery. The findings deepen our understanding of the mechanisms underlying minocycline-mediated neuroprotection in AIS.

endovascular therapy can not improve disability-free survival after AIS compared with IV-tPA, indicating that infarcted tissue is technically successful recanalization but no longer salvageable [4,5]. Thus, medicines for effective stroke recovery are likely to remain an essential part of stroke care in the future.
Post-stroke in ammation plays an essential role in the survival and regeneration of nerve cells after AIS [6,7]. The lack of oxygen and energy during AIS leads to rapidly damaging nervous tissue and activates the innate immune system to support healing, including clearing the dead tissue. Brain-resident microglia and in ltrating peripheral leukocytes migrate to the infarct area and stimulate the production and secretion of pro-in ammatory cytokines. These immune cells contribute to the clearance of cellular debris and dead cells from the injured brain through phagocytosis. However, they also destroy salvageable tissue, which is only damaged by the lack of oxygen and energy not yet dead. Thus, post-stroke in ammation is both detrimental and bene cial at different stages [8]. The proper function of these immune cells is fundamental to post-stroke recovery.
Microglia mediates the in ammatory cascades as the rst defensive line and modulates brain repair after stroke [8,9]. Increasing evidence suggests microglia could experience two distinct phenotypes after activation: a "classical" deleterious M1 phenotype stimulated by lipopolysaccharide (LPS) and interferonγ (IFN-γ) and an "alternative" neuroprotective M2 phenotype stimulated by interleukin-4 (IL-4) [10]. After AIS, M1 microglia promotes in ammation and aggravates brain damage by releasing pro-in ammatory mediators at the injury site where the repair process of M2 microglia is depressed. Thus, promoting the phenotype shift of M1 to M2 has become an adjunct therapy in recovery after stroke [11,12].
Minocycline, an anti-infective agent of the tetracycline family, was known for its anti-in ammatory effects in neurological disorders [13][14][15], including AIS [16]. We previously found that minocycline could prevent microglia activation and inhibit the NLRP3 in ammasome up to 72 h after the transient middle cerebral artery occlusion -reperfusion (MCAO/R) [17]. Since it is reported that minocycline can improve functional recovery after cerebral ischemia [18,19], we have a hypothesis that minocycline may shift microglia from M1 to M2 phenotype in recovery after induction of MCAO/R.

Animals and treatment
Male C57BL/6 mice (aged 10-12 weeks, weighing 23-25 g) were purchased from the Animal Core Facility of Nanjing Medical University. All mice were housed in ve per cage groups in a standardized light-dark cycle at 22 °C and given free access to food and water. All experimental procedures were approved by the Institutional Animal Care and Use Committee of Nanjing Medical University. All procedures were designed to minimize suffering and reduce the number of animals used. MCAO/R model MCAO/R mice were established as previously described [17]. For the transient model, reperfusion was produced by withdrawal of the 6-0 nylon lament (Doccol Corp., Readlands, CA, USA) 60 min after the occlusion. In the sham group, arteries were visualized but not ligated. After the surgery, body temperature was controlled with a heating pad and kept at 37 ° C until mice wholly recovered from anesthesia. The mice were then returned to their home cages.

Grouping and treatment
After 24 h of reperfusion, the model mice were subjected to neurological function evaluation based on the neurological severity scores (NSS) system. Mice with NSS scores of 4 to 10 were randomly divided into 4 groups and received intraperitoneal injections with saline, minocycline 10 mg/kg, 20 mg/kg, or 50 mg/kg. They were then sacri ced on days 3, 7, or 14 after reperfusion according to different experimental purposes. Minocycline (Sigma-Aldrich, St. Louis, MO, USA) was freshly diluted to the concentrations needed with saline.

Neurobehavioral function evaluation
Using the revised NSS system [20], the neurobehavioral evaluation of mice was examined by two trained observers who were blinded to the drug treatment group. According to the NSS presented in Table 1, mice's neurological status was examined on day 1 and day 14 following MCAO/R (Fig. 1a). The 1-day assessment was used to verify the ischemic state of MCAO/R groups. The 14-day assessment was used to evaluate motor and behavioral impairments in recovery after induction of MCAO/R.

Rotarod test
We carried out the rotarod test to calculate mice's motor ability as previously described [21]. All mice were acclimatized to the rotarod by undergoing a 3-day training program before MCAO/R surgery experiments. Only mice that can remain on the rotarod for 180 s were used for the experiments. On day 14 following MCAO/R, mice performed three trials on the rotarod, with an interval of 5 min. The mean of the three trials was used to show the functional recovery of mice.

Corner turning test
The corner test was performed as previously described [22]. Brie y, a mouse was left in the test device consisting of two vertical boards at a 30° angle. When the mouse entered the corner, the direction in which it contacted the board with the vibrissae was recorded. The non-ischemic mouse turned left or right with equal frequency, but the mouse suffered from MCAO/R preferentially turned toward the impaired (right side in our experiment). The number of turns in each direction was summed from 10 trials. The following formula was used to calculate right-biased turning percentage: right-biased turning percentage = right-biased turning / total turning × 100%. TTC assay As previously described [21], the cerebral infarct volume of nine mice in each group was measured with 1% TTC (dissolved in saline) on day 7 following MCAO/R. The infarcted regions were acquired by a digital camera and quanti ed using ImageJ software (NIH, Bethesda, MD, USA). The following formula was used to calculate infarct percentage: Infarct percentage = infarct volume/volume of the contralateral hemisphere × 100%.

Immunohistochemistry and cell counting
On day 3 or day 7 following MCAO/R, six mice were anesthetized with pentobarbital sodium (80 mg/kg) and then perfused with saline and 4% paraformaldehyde as previously described [21]. Their brains were xed overnight in 4% paraformaldehyde at 4 °C, followed by dehydration in 30% w/v sucrose at 4 °C. The frozen coronal sections of 30 μm were cut with a vibrating microtome (Leica CM1950, Nussloch, Germany). VT, USA) as described previously [23].
Quantitative real-time polymerase chain reaction Quantitative real-time polymerase chain reaction (qRT-PCR) was performed as previously described [24].
Cell culture and treatment Primary microglia from cerebral cortex of newborn mice were cultured and isolated as previously described [25]. At days in vitro (DIV) 10, microglia cells were seeded on 6-well plates and incubated with overnight at 4°C. The membranes were washed and incubated with corresponding secondary antibodies for 1 h at room temperature. Finally, the membrane was observed and analyzed with Tanon 5200 (Tanon Science and Technology Co. Ltd, Shanghai, China).
Neuron-microglia co-culture Primary microglia and cortical neurons were co-cultured in a two-layer system (Transwell, Corning, Corning, NY, USA). Brie y, microglia on Transwell inserts were treated with 100 ng/mL LPS plus 20 ng/mL IFN-γ followed by 50 μM minocycline or saline for 24 h. Then, the medium was removed and washed with fresh medium three times. Primary cortical neurons were cultured and subjected to oxygen-glucose deprivation/reperfusion (OGD/R) for 1 h, as previously described [24]. When the medium was replaced with a normal medium, the microglia inserts were added on top of neuronal cultures to generate a neuronmicroglia co-cultures. After 24 h of co-culture, neuronal cell death and survival were measured by thiazolyl blue tetrazolium bromide (MTT; Sigma-Aldrich, St. Louis, MO, USA) assay and lactate dehydrogenase (LDH) assays (Nanjing Jiancheng Bioengineering Institute, Nanjing, China) according to the manufacturer's protocol.

Statistical analysis
All data were expressed as mean ± S.E.M. Statistical analysis was performed using GraphPad Prism software (San Diego, CA, USA). A log-rank test was used to compare the survival rate of the MCAO/R model. A non-parametric Kruskal-Wallis H-test was used to compare the NSS scores of the MCAO/R model. Other parametric data were analyzed via one-way analysis of variance (ANOVA) followed by Tukey's test or a two-sample t-test. P < 0.05 was considered statistically signi cant.

Results
Minocycline enhanced functional recovery after MCAO/R model To con rm the effects of minocycline on functional recovery following AIS, we treated mice subjected to MCAO/R with saline or different minocycline concentration (10, 25, or 50 mg/kg, i.p., daily for 2 wk) at 24 h after reperfusion (Fig. 1a). No mice died during the surgery. During the rst 1 wk after reperfusion, 18 mice in MCAO/R group, 11 mice in MCAO/R + 10 mg/kg minocycline group, 8 mice in MCAO/R + 20 mg/kg minocycline group, and 3 mice in MCAO/R + 50 mg/kg minocycline group died. Since mice that survived did not die during the time window of 14 days, the survival rate was 45.5%, 56.0%, 68.0%, and 88.0%, respectively (Fig. 1b). The survival rate was signi cantly different among groups (P = 0.0019), and between MCAO/R group and MCAO/R + 50 mg/kg minocycline group (P = 0.0016).
To evaluate the initial neurological de cit after MCAO/R and monitor neurobehavior on day14 after minocycline treatment, we rated each mouse's performance suffered from MCAO/R according to NSS scores. The mice suffered from MCAO/R showed high neurological de cit scores at d 1 after reperfusion ( Fig. 1c, open column). There was no difference among the four groups (P = 0.9716). At day 14 after minocycline treatment, three MCAO/R + minocycline groups demonstrated less neurological impairment compared to MCAO/R + saline group, and the mice treated with 25 mg/kg (P = 0.0001) and 50 mg/kg (P < 0.0001) minocycline showed a signi cant difference (Fig. 1c). Consistent with the NSS scores, the mice in MCAO/R + minocycline groups showed increased latency time in the rotarod test (Fig. 1d). The latency time was signi cantly different among groups (F (4, 77) = 42.83, P < 0.0001), between MCAO/R + saline group and MCAO/R + 25 mg/kg minocycline group (P = 0.0003), and between MCAO/R + saline group and MCAO/R + 50 mg/kg minocycline group (P < 0.0001). Meanwhile, the mice in MCAO/R + minocycline groups also showed reduced right-biased turning in corner turning test (Fig. 1e). The percentage of rightbiased turning was signi cantly different among groups (F (4, 77) = 41.41, P < 0.0001), between MCAO/R + saline group and MCAO/R + 25 mg/kg minocycline group (P < 0.0001), and between MCAO/R + saline group and MCAO/R + 50 mg/kg minocycline group (P < 0.0001). These results indicate that minocycline improves functional motor recovery at d 14 following MCAO/R. Minocycline alleviated neuronal injury on day 7 following MCAO/R Since mice did die during the time window of the rst 7 days, we chose day 7 after reperfusion to observe the effects of minocycline on neuronal injury after MCAO/R (Fig. 2a).

Minocycline prevented neurons from OGD/R-induced cell death in neuron-microglia co-cultures
To clarify the neuroprotective role of modulating microglia polarization, we directly added primary cortical microglia to co-cultures of primary neurons (Fig. 7a). The post-OGD/R neurons were incubated with the different groups of primary microglia. The neurons from minocycline-pretreated M1 microglia with LPS plus IFN-γ attenuated OGD/R-induced neuron damage compared to the M1 microglia with LPS plus IFN-γ, revealed by increasing cell viability (P < 0.0001; Fig. 7b) and reduced LDH release (P = 0.0019; Fig. 7c).
The results show that minocycline ameliorates neuronal cell death via modulating microglial polarization, contributing to brain functional recovery after ischemic stroke.

Discussion
Numerous studies have shown the neuroprotective effects of minocycline in ischemic stroke. In particular, both experimental [28,29] and clinical evidence [30][31][32] indicate that repeated minocycline treatment can improve functional outcomes after AIS. Previous studies in our laboratory's experimental stroke have shown that minocycline suppresses microglial activation via regulating NLRP3 in ammasome activation [17] and promotes BDNF expression via inhibiting miR-155-mediated repression [24]. Here, using the MCAO/R model, we con rmed that a two-week treatment with minocycline could increase the survival rate and promoted functional outcomes through alleviating neuronal injury and reactive gliosis. We further observed an additive effect of minocycline on microglia polarization to the M1 and M2 phenotypes in vitro and in vivo. Thus, we demonstrated for the rst time that minocycline promoted microglial M2 polarization and inhibited M1 polarization, leading to neuronal survival and neurological functional recovery.
Microglia activation and monocyte in ltration normally peak 48 -72 h after stroke onset and last for several weeks [33]. In the present study, since the mice's deaths occurred in the rst week, we observed the involvement of microglia in ischemic injury on 3 d and 7 d after MCAO/R. We found the proliferation and activation of Iba-1 + cells (macrophage/microglia) in the ipsilateral cortex and striatum after MCAO/R. Interestingly, most activated microglia were present around the lesion site on day 3 and then in ltrated the ischemic core on day 7, indicating that microglia may clear the cell debris by phagocytosis crucially in the resolution of in ammation. As previously reported [26,33], both CD16/32 and CD206 were detected in the mouse brain on day 7 after MCAO/R, while both M1 markers (IL-1β, IL-6, iNOS, and TNF-α) and M2 markers (Arg1, IL-10, TGF-β, and Ym1) were observed. These results demonstrated an M1/M2 polarization microglia/macrophages in the ipsilateral hemisphere of MCAO/R mice. Furthermore, we found that the number of M1 microglia in the penumbral was much larger than the number of M2 microglia on day 7 after MCAO/R. A previous study reported an M2-to-M1 transition in microglia/macrophages in the ischemic core [34]. During the rst two weeks after the onset of ischemic stroke, M1 microglia has increased in the ischemic core over time, whereas M2 microglia arrives at the peak at day 5 and then intensively decreased throughout the second week. The long-term activation of M1 microglia can lead to secondary neuronal damage in the next few months. Our results were consistent with these ndings and further indicated that minocycline reverses this M2-to-M1 transition. The experiment in vitro further revealed that minocycline reduced microglial M1 polarization and promoted M2 polarization. We also found that microglia treated with minocycline prevented OGD/Rinduced neuronal cell death in the neuron-microglia co-cultures. Thus, one novel nding in the present study is that minocycline reversed ischemic-induced M2-to-M1 transition and alleviated neuronal injury in experimental stroke, suggesting that minocycline shifts microglia from M1 to M2 phenotype to improve functional recovery after cerebral ischemia.
Numerous studies reported a critical role of minocycline in the suppression of microglial activation. Minocycline could selectively attenuate M1 polarization of microglia via inhibiting the upregulation of NF-κB in the LPS-stimulated microglia and SOD1(G93A) mice [35]. Chronic administration of minocycline modulates the gene expression of M1-M2 microglial to attenuate neuropathic pain behavior in a rat model of depression [36]. Recently, the results of a single systemic injection of LPS showed that minocycline inhibited neuroin ammatory responses and promoted M2 polarization of microglia through activation SIRT1 [37]. However, the underlying mechanism of minocycline in regulating microglia polarization after stroke is unknown. The Janus kinase (JAK)/STAT signaling pathway is a crucial mediator of neuroin ammation and plays an important role in the progression and pathogenesis of ischemic stroke [38]. Both STAT1 and STAT6 contribute to an ischemia-induced switch for microglial activation [38]. Microglia can be polarized to M1 phenotype via STAT1 activation while to M2 phenotype via STAT6 activation. In the present study, LPS plus IFN-γ induced the M1 microglia polarization, resulting in the STAT1 activation and increased IFN-γ-dependent gene for M1 polarization such as IL-1β, IL-6, iNOS, and TNF-α. Previous studies demonstrate that IFN-γ also suppresses the recruitment of STAT6 to the IL-4R and inhibits its phosphorylation in Th1 cells [39]. We only observed the IFN-γ-induced decreased Arg1 gene, not other M2-related genes or the phosphorylation of STAT6 in vivo and in vitro. In the present study, minocycline suppressed the phosphorylation of STAT1 and activated the phosphorylation of STAT6, indicating that minocycline regulates M1/M2 microglia polarization via the STAT1/STAT6 pathway.

Conclusion
We demonstrated for the rst time that minocycline improved functional recovery after cerebral ischemia via shifting microglia from M1 to M2 phenotype. We found that minocycline reduced the production of genes for M1 polarization and enhanced the expression of genes for M2 polarization through regulating the STAT1 and STAT6 signaling. It would be an important topic that requires further investigation on how minocycline regulates the balance of STAT1 and STAT6 activation in microglia after cerebral ischemia.

Consent for publication
Not applicable.

Availability of data and materials
Not applicable.

Competing interests
The authors declare that they have no competing interests.