The present study was performed in strict accordance with the guidelines of the Institute of Biochemistry and Cell Biology, Chinese Academy of Sciences. All experimental protocols were approved and overseen by the Animal Care and Use Committee of the Shanghai Institute of Biochemistry and Cell Biology, Chinese Academy of Sciences. All mice were maintained in pathogen-free conditions.

BV2 cells were cultured and maintained in Dulbecco's Minimal Essential Medium (DMEM), supplemented with 10% fetal bovine serum (FBS) and 100 U/ml penicillin and 0.1 mg/ml streptomycin. Primary microglia cells were prepared from wild type SD rat on postnatal day 1. In brief, the mixed rat brain region including cortex and hippocampus was dissected and the meninges were carefully removed. The brain tissues were dissociated into single cells by gentle scissoring and pipetting. The resultant cell suspension was seeded to a T75 flask with DMEM supplemented with 10% FBS and 100 U/ml penicillin and 0.1 mg/ml streptomycin. Microglial cells were isolated from the astrocyte monolayer sheet by shaking.

In all cell-based assays, BV2 and primary microglia cells were pre-treated with α-asarone (Fluka) at various concentrations (0.3–100 μM) for 2 h, followed by 0.1 or 1 μg/ml LPS (055:B5, Sigma) stimulation for further 24 h.

BV2 cells and cultured rat microglia cells were fixed with 4% PFA in PBS for 15 min at room temperature, followed by 30 min blocking in PBS containing 1% BSA and 0.1% Triton X-100. The cultured rat microglia cells were incubated with anti-Iba1 (1:1000, WAKO) primary antibody for 2 h at room temperature, and Cy3-conjugated anti-rabbit IgG (1:1000) secondary antibody for 1 h avoiding light exposure. BV2 cells were incubated with PE-conjugated CD86 (1:100) or APC-conjugated CD80 (1:100) antibodies overnight at 4°C, avoiding light exposure. Afterwards, cells were stained with DAPI (1:1500) for 10 min at room temperature.

The cytotoxicity potential of α-asarone were examined in the presence and absence of LPS. BV2 cell viability was tested using the CellTiter-Glo luminescent cell viability assay (Promega) following the manufacturer's instructions.

BV2 and primary microglial cells were grown in 48-well plates till 80% confluent. The monolayer of cells was then wounded with a sterile 200 μl pipette tip, and washed three times with sterile PBS. Afterwards, the cells were incubated with α-asarone (0.3, 3, and 100 μM) in the presence and absence of LPS for 24 h. Additionally, primary microglia cells were stained with CellTrace™ CFSE Cell Proliferation Kit (Invitrogen C34554) before microscopic imaging.

BV2 cells were seeded into 24-well plates at appropriate densities. The LPS and α-asarone treatments were applied according to Section Cell Culture and Treatment. The MCP-1 (Peprotech), CCR2 antagonist RS504393 (Sigma) and mouse biotinylated MCP-1 antibody (ELISA, DAKEWE Biotech Company) were applied with LPS where necessary. At the end of 24-h treatment, cell images were captured by Zeiss Observer Z1 microscope with 20X objective (Plan-Neofluar 20X/0.4). The BV2 immunocytochemistry staining images were documented using Olympus FV1200MFE microscope with UPlanSApo 40X2 (40X/0.95) objective. Primary microglia immunocytochemistry staining images were captured using Olympus IX51 camera with 20X/0.45 Olympus objective. In the Scratch assay, BV2 cells and primary rat microglia cells migrated into the open area was imaged by Zeiss Observer Z1 microscope with Zeiss Plan-Neofluar 5X and 20X objectives (Plan-Neofluar 5X/0.16, Plan-Neofluar 20X/0.4).

For morphological analysis, only the single distinct cells were selected. Dividing cells and cells attached to others were not included. Image data were processed by Fiji ImageJ software (Version 2.0.0). The following parameters were quantified: cell area, perimeter, Feret's diameter (the longest distance between any two points along the selection boundary) and circularity (4πAreaPerimeter2). For BV2 cell CD86 and CD80 expressions, the integrated intensities of each single distinct cell were measured, the MFI (mean fluorescent intensity) levels were calculated with respect to the “Short” and “Long” morphologies.

For ELISA assays, the supernatants from each treatment were collected and stored at −20°C if not used immediately. Prior to the assay, the supernatant was centrifuged to remove cellular debris, and diluted by an appropriate factor. IL-6 and MCP-1 expression were determined using mouse-specific pre-coated ELISA kits (DAKEWE Biotech Company) according to manufacturer's instructions.

Total RNA was extracted from BV2 cells using the TRI Reagent® (Sigma) following the manufacturer' s instructions. Purity and integrity of the RNA was assessed by NanoDrop 1000 Spectrophotometer (Thermo Scientific) and stored at −80°C if not for immediate use. The first-strand cDNS synthesis was performed using the TIANScript M-MLV kit (TIANGEN) following manufacturer's instruction. The synthesis was carried out with the addition of rRNasin® (Recombinant rRNasin® Ribonuclease Inhibitor, Promega).

The mRNAs expression was determined by quantitative real-time PCR using the 2x HotStart SYBR Green qPCR Mater Mix kit (ExCell). Each reaction was performed in duplicates, and the supermix contained 4 μl pre-diluted cDNA and 0.25 μM primers in a total volume of 25 μl. The reaction parameters were as follow: 10-min of a 95°C cycle following 40 cycles of denaturation (95°C for 30 s), annealing (60°C for 30 s) and extension (72°C for 30 s). An additional cycle was performed for evaluation of primer's dissociation curve at 95°C for 1 min, 60°C for 30 s and 95°C for 30 s. Primer sequences used in the experiments are listed in the Supplementary Table 1.

BV2 cells were seeded into the 96-well glass bottom black plate. After 18 h, cells were treated with LPS and α-asarone according to Section Cell Culture and Treatment. The plate was imaged using the Cellomics ArrayScan VTI 700 (Thermo Scientific, Pittsburgh, PA, USA) with 20X objective under bright field. For each well, real-time images were captured from 16 adjacent fields at each time point, at 10-minute intervals for 24 h. During the course, cells were maintained at 37°C with 5% CO₂. For data processing, each single distinct cell was selected. The long cell number, motility, cell area and circularity parameters were analyzed using Fiji ImageJ (version 2.0.0).

Statistical analysis was performed using GraphPad Prism 6 software (Graphpad Software, La Jolla, CA, USA). Results were analyzed by unpaired two-tailed Student's t-test to determine the significance of treatment sets. For group comparisons, two-way ANOVA analysis of variance was performed. All data were presented as mean ± SEM. P < 0.05 was considered to have significant difference.

Previously, α-asarone was reported to promote neuronal progenitor cells (NPC) proliferation (Mao et al., 2015). We first confirmed that α-asarone (0.3–100 μM) did not promote BV2 microglia proliferation in the presence and absence of LPS stimulation (Supplementary Figure 1). The results also indicated that α-asarone did not exhibit cytotoxicity. Therefore, we proceeded to characterize the effect of α-asarone on microglia morphology. We examined the BV2 cell morphological changes at basal and LPS-stimulated conditions. First, we noticed the cell morphology changed with respect to differential treatment conditions (Figure 1A). We defined the “short” and “long” cells and recorded the proportion of each morphological type within the cell population. The results showed majority of cells under untreated control and α-asarone treatment alone conditions stayed short (Figure 1B). Stimulation by LPS greatly increased the number of “long” cells. However, the addition of α-asarone significantly attenuated such morphological changes in a concentration-dependent manner. We then further quantitatively defined the morphology using four parameters: the cell area, perimeter, Feret's diameter and circularity (Figures 1C–F). The results were consistent with the Figure 1B findings. Under the conditions when cells were stimulated with LPS, α-asarone treatment significantly reduced cell area, perimeter and Feret's diameter and circularity, also in a dose-dependent manner. From this, we raised the question whether these long cells correspond to the activated microglia. By immunocytochemistry assay, we examined the CD86 and CD80 expression in “long” and “short” cells after LPS and α-asarone treatments as above (Figure 1G). We found significant increases in both CD86 and CD80 expression in “long” cells compared with “short” cells, indicating that the “long” cells, indeed, correspond to the activated microglia. Therefore, these data implied that α-asarone could attenuate activated microglia morphological changes.

In addition, we analyzed the morphological changes of activated BV2 and primary microglia using a wound-healing model. As previously described, LPS induces microglia migration toward a scratched wound (Karlstetter et al., 2010, 2011; Lively and Schlichter, 2013). We created a single uniform scratch on the monolayer of cells (Supplementary Figures 2A, 3A). After migrating toward the open wound, the morphology of each distinct cell within the wounded area was analyzed. We also determined the proportion of “long” and “short” cells (Supplementary Figure 2B), and quantitatively defined the cell area, perimeter, Feret's diameter and circularity (Supplementary Figures 2C–F, 3B–E). The results were consistent with Figure 1 findings. For primary cultured microglia, no change in cell circularity was detected.

We further examined the effect of α-asarone on BV2 microglia morphological dynamics on a 24-h time course. BV2 cells were pre-treated with 3 μM α-asarone in the presence of LPS or with 100 μM α-asarone alone. Micrographic images were documented at 10-min intervals for 24 h. For each well of a specific treatment condition, we captured images from 16 adjacent fields at each time point. Only data for typical time points were shown in Figure 2. Stimulation by LPS resulted in an increasing trend of long cell numbers. The attenuation effect of α-asarone was displayed from 7 h onwards (Figure 2B). The differences between the two treatment conditions were most significant at 24 h. This was consistent with earlier findings (Figure 1). We also determined the cell area and circularity parameters (Figures 2C,E). Changes in both parameters were monitored for 1 h and the results were normalized to the average of “control.” Only representative data at four specific time-phases were shown. At 23–24 h time-phase, we detected reduction in cell area and an increase in the circularity ratio related to the effect of α-asarone on the LPS-activated cells. In addition, at 5–6 h time phase, α-asarone showed a tendency to reduce cell area as well as to increase cell circularity in comparison to LPS stimulation alone. However, no statistical significance was detected. There was also no significant difference in motility, which may due to large variance in the data (Figure 2D). Time-lapse imaging consistently showed that α-asarone attenuated LPS-induced microglia morphological changes at 3 μM. A representative video was provided in the supplementary material (Supplementary Video), on the left panel cells were stimulated with LPS, and the right panel the LPS-activated cells were incubated in the presence of 3 μM α-asarone.

Previous study reported that between 50 and 250 μM, α-asarone exhibited anti-inflammatory effect on the LPS-stimulated microglia cells (Kim et al., 2015). Nevertheless, as demonstrated in Figure 1, the minimum effective concentration of α-asarone required to modulate the morphology of activated microglia was 3 μM, which is significantly lower compared to the previous report. In order to determine whether the morphological modulations of the activated microglia were associated to the attenuation of inflammatory responses, we analyzed the expression of various pro-inflammatory cytokines. We found that 3 μM α-asarone did not inhibit the pro-inflammatory cytokine IL-1β and IL-6 mRNA expressions (Figures 3A,B). Protein level detection by ELISA showed consistent results. Increasing α-asarone concentration (above 30 μM) indeed resulted in the inhibition of LPS-stimulated IL-6 expression (Figure 3F). In addition, we confirmed these findings using 0.1 μg/ml LPS as in the previous report (Kim et al., 2015) (Supplementary Figure 4). Our results were also consistent with the previous study, such that at high concentrations (100 and 300 μM) α-asarone significantly reduced the LPS-induced IL-1β and IL-6 expressions. Meanwhile, lower concentrations of α-asarone (1 and 3 μM) did not affect IL-1β and IL-6 levels, consistent with Figures 3A,B. These results implied that the morphological modulation of activated microglia might not associate with the attenuation of inflammatory responses.

In order to define a mechanism by which α-asarone modulated microglia morphological dynamics, we analyzed the expression of MCP-1, MIP-1α and ROCKII, which were reported to influence cell migration and morphology (Xu et al., 1996; Lentzsch et al., 2003; Widera et al., 2004; Yoneda et al., 2012). The results showed that MCP-1 expression was reduced by α-asarone in a concentration-dependent manner (Figures 3C,G). The inhibition of LPS-induced MCP-1 accumulation was observed when exposed to both low (1–3 μM) (Figures 3C,G, 3 μM) and higher concentrations (10–100 μM) (3G, 10, 30, and 100 μM) of α-asarone. On the other hand, no significant changes in MIP-1α and ROCKII expressions where detected (Figures 3D,E). Therefore, MCP-1 was associated with the LPS-induced morphological changes, and α-asarone may attenuate such morphological change by regulation of MCP-1 expression. By using the CCR₂ antagonist (10 μM RS504393) or MCP-1 antibody, we confirmed on the inhibitions of LPS-induced morphological changes (Figures 3H–K).

Inversely, the addition of MCP-1 (100 ng/ml) retained the LPS-resulted microglia morphological change in the presence of α-asarone. We recorded an increase in the long cell numbers associated with MCP-1 even in the presence of α-asarone treatment (Figure 3L). Concomitantly, changes in cell area, perimeter, Feret's diameter and circularity parameters together demonstrated an opposing effect of MCP-1 to α-asarone (Figures 3M–P). These results served clear indications that the attenuation by α-asarone to LPS-induced morphological changes was effectively rescued by MCP-1.

In order to validate our findings, we examined the morphological changes in primary rat microglia. We first confirmed the α-asarone effect on IL-1β, IL-6, and MCP-1 expressions (Figures 4A–C). The results showed that at higher concentrations (100 and 300 μM), α-asarone inhibited the accumulation of LPS-induced IL-1β, IL-6, and MCP-1 expressions. At lower concentration (3 μM), α-asarone specifically reduced the LPS-induced MCP-1 expression. These results correlated with the findings in BV2 cells. Subsequent morphological analysis revealed that untreated microglial cells maintained relatively small somata sizes, as defined by the cell area (Figures 4D,E). Stimulation by LPS resulted in a significant enlargement in cell area, and the reverse effect was observed when α-asarone was applied. Nevertheless, α-asarone attenuation on the LPS-induced somata enlargement was effectively rescued by MCP-1 (Figures 4D,E). We also detected significant increase in cell perimeter and Feret's diameter in the presence of LPS stimulation (Figures 4F,G). In addition, co-treatment with α-asarone showed a decreasing trend in both parameters, but the difference for Feret's diameter measurement was not significant. In addition MCP-1 rescued cell perimeter and Feret's diameter in the presence of α-asarone. Finally, no significant changes in cell circularity were detected under all conditions (Figure 4H). These results confirmed that α-asarone attenuation on LPS-induced morphological changes was rescued by MCP-1.

To determine the role of α-asarone in microglia in vivo, transgenic line Tg(Apo-E:eGFP) with microglia in green was used (Li et al., 2012). First of all, zebrafish embryos were treated with 0.3–3 μM α-asarone from 12 hpf to 5 dpf, and no developmental defects were observed under these concentrations (Supplementary Figure 5A). Therefore, we treated zebrafish embryos with 3 μM α-asarone in the presence and absence of E coli from 4 dpf for 24 h, and examined the morphology and dynamics of only resting microglia. The in vivo time-lapse imaging was carried out at 5-min intervals for 60 min according to previous study (Li et al., 2016). No significant difference was observed in the resting microglia cell area, deformation speed and tip number (Supplementary Figures 5B–G). Subsequently, we analyzed the activated microglia dynamics by injection of the red-fluorescent E. coli into the gap at the optic tectum region (Figure 5A), and images were then captured (Figure 5B). Following the previous description for microglia states (Kettenmann et al., 2011), we examined the percentage of active-state microglia in the presence of α-asarone. The results showed a remarkable decrease (control vs. α-asarone 79.4 ± 5.3 vs. 59.6 ± 6.9, p < 0.05, Figure 5C). Quantitative analysis of active microglia morphological dynamics showed that the effect of α-asarone to the average velocity of microglia cells was not statistically significant. However, the effect in the overall trend was significant, which indicated α-asarone maintained the activated microglial migration velocity (p < 0.01, Figures 5D,E). Further analysis revealed that α-asarone treatment led to approximately 50% decrease in the average tip number (asarone 53% ± 7% of control, p < 0.05, Figure 5G) without changes in the cell area (asarone 105% ± 4% of control, p = 0.55, Figures 5H,I). In addition, the dynamical changes of tip numbers change over the 60-minute time course also showed significant difference with relation to α-asarone treatment (p < 0.0001, Figure 5F). These results suggested that α-asarone might affect the microglia activation and function without modulating the resting state dynamics.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.