Anti-inflammatory and immunomodulatory mechanisms of atorvastatin in a murine model of traumatic brain injury

Male C57BL/6 mice, 6–8 weeks old (20–23 g), were purchased from the Experimental Animal Laboratories of the Academy of Military Medical Sciences (Beijing, China). The experimental protocols for this study were approved by the Tianjin Medical University Animal Ethics Committee (Tianjin, China). All mice were maintained with free access to food and water in a temperature-controlled (20 ± 2 °C) and humidity-controlled (55 ± 5%) vivarium under a 12 h light/dark cycle, and were adapted to the environment for a week before the experiments.

A total of 150 mice were randomly assigned to the following four groups: sham + saline group (n = 30); sham +1 mg/kg/day atorvastatin group (n = 30); TBI + saline group (n = 45); TBI + 1 mg/kg/day atorvastatin group (n = 45). Post-treatment assessments are shown as a schematic in Fig. 1a. Proper measures were taken to minimize the number of mice used and the pain or discomfort they might experience.

To determine the efficacy of atorvastatin (Atorvastatin calcium tablet, Pfizer Inc.; dissolved in 0.9% saline) on neuroinflammation in the acute phase of TBI, the initial dose was administered orally by intragastric gavage (i.g.) as early as 1 h after TBI [26]. A pilot dose-response analysis was performed using three different doses (1, 5, and 10 mg/kg/day) to determine the optimal dose of atorvastatin, and neurological deficits were evaluated as main outcomes. The results demonstrated that mice treated with atorvastatin performed significantly better than those treated with saline; however, there was no significant difference among the three different doses at all tested time points. Therefore, a dose of 1 mg/kg/day (with the minimal side effect) was chosen for all subsequent experiments. Mice in sham + saline and TBI + saline groups received equal volumes of 0.9% saline with the same schedule.

TBI was induced using a controlled cortical impact (CCI) device (eCCI-6.3 device, Custom Design & Fabrication, USA). Briefly, mice were anesthetized with 10% chloral hydrate through intraperitoneal (i.p.) injection (3 mL/kg) and then positioned in a stereotaxic frame with a heating pad to maintain body temperature. After the scalp and fascia were retracted, a 3.5-mm diameter hole was drilled on the right cerebral hemisphere (2.0 mm posterior from bregma and 2.0 mm lateral to the sagittal suture) to expose the intact dura. For moderate-TBI induction, a 3-mm-flat impactor tip was used to impact the exposed dual (impact parameters: velocity: 4.5 m/s, depth: 2 mm, dwell time: 200 ms). After CCI injury, the scalp incision was sutured, and the mice were placed in a heating pad until anesthesia recovery. Sham-injured mice received all of these procedures, but did not receive a CCI.

At each time point, mice were anesthetized deeply using 10% chloral hydrate (3 mL/kg, i.p.), the spleens were surgically removed and mononuclear cells were isolated using RBC lysis buffer (eBioscience, San Diego, CA, USA). The anesthetized mice were then perfused intracardially with ice-cold phosphate-buffered saline (PBS). Brains were isolated rapidly and were used as follows. (1) In flow cytometry analysis, brain mononuclear cells were isolated via a discontinous percoll gradient method. Briefly, cerebral hemispheres were gently pressed through a 40-μm nylon cell strainer (BD Bioscience, Franklin Lakes, NJ, USA), the tissue suspension was then resuspended in 5 mL of 30% percoll (Sigma-Aldrich, St Louis, MO, USA), and slowly layered on top of a 5 mL of 70% percoll followed by centrifugation at 500×g for 20 min at 18 °C. Mononuclear cells were collected in the 30–70% interphase, washed twice with PBS, and resuspended in 100 μL of flow cytometry staining buffer (eBioscience) for further use. (2) For immunostaining analysis, fresh brains were fixed in 4% paraformaldehyde for 24 h at room temperature and were then embedded in paraffin (for immunohistochemistry) or the Tissue-Tek O.C.T. compound (dehydration by 20% and then 30% sucrose before tissue embedding, for immunofluorescence). Subsequently, a 3-mm thick coronal block, spanning the entire injured cortex, was harvested, and transverse 6-μm thick sections were then cut using a microtome (Leica, Germany). 3) In enzyme-linked immunosorbent assay (ELISA) and quantitative real-time PCR (qRT-PCR) analysis, the peri-contusional cortex (see Fig. 1b) in TBI mice and the equal area in sham-injured mice were dissected on ice, and then immediately frozen in liquid nitrogen. Some of which were homogenized in nine volumes (1:9, w/v) of ice-cold saline, the supernatants were collected after centrifugation at 1000×g for 15 min; the others were immediately immersed in RNA stabilization solution (Invitrogen, Carlsbad, CA, USA) to preserve the RNA quality and quantity.

Neurological deficits were assessed using well-established modified neurological severity score (mNSS) and Rota-rod at baseline, 24 and 72 h after TBI or sham surgery by two investigators who were blinded to the experimental design. Each behavioral test was repeated twice with four different trials to validate the data.

The mNSS test consists of ten different tasks that can evaluate the motor (muscle status, abnormal movement), sensory (visual, tactile and proprioceptive), balance, and reflex functions of mice. Neurological function was graded from 0 to 18 (0 = normal function; 18 = maximal deficit). One point was scored for each abnormal behavior or for the lack of a tested reflex. Therefore, higher scores implying greater neurological injury.

The fine motor coordination and learning were assessed using an accelerating Rota-rod apparatus (RWD Life Science, Shenzhen, China). On the day before injury, mice were trained on the Rota-rod for three consecutive trials at a slow rotational speed (4 rpm/min) for 1 min to adapt to the rod, followed by four additional trials with an accelerating rotational speed (from 4 to 40 rpm in 5 min) to obtain baseline latency. On each testing day, the mice were given four 300-s accelerating Rota-rod trials with an inter-trial interval of 30 min. The average latency to the first fall off the rod was recorded. Passive rotation, accompanying the rotating rod without walking, was also considered as a fall.

For flow cytometry analysis of cellular components in the injured brain, the isolated cells were stained with fluorescently labeled antibodies: CD45-FITC, CD11b-APC, Ly6G-PE, CD3-PE-Cy7, CD45R (B220)-APC, NK1.1-PE, CD206-PE, CD86-PE-Cy7 and the appropriate isotype control according to the manufacturer’s protocols (eBioscience). To detect Tregs in the brain and spleen, the isolated cells were stained using a commercial Mouse Tregs Staining Kit (eBioscience). Briefly, cells were surface stained with anti-mouse CD4-FITC and anti-mouse CD25-PE for 30 min at 4 °C, fixed, and then permeabilized overnight at 4 °C using the fixation/permeabilization solution and subsequently stained with the anti-mouse/rat Foxp3-APC antibody. Flow cytometry was performed on a FACS Aria III apparatus (BD Bioscience) and the obtained data were analyzed by Flow Jo software 7.6.1(Tree Star, US).

Paraffin-embedded sections were deparaffinized with gradient ethanol and xylene, and then boiled in a microwave with citrate buffer (pH 6.0) for 30 min to retrieve antigens. Following three washes with PBS, sections were incubated with 3% hydrogen peroxide and 3% bovine serum albumin (BSA, Sigma-Aldrich) to block endogenous peroxidase and nonspecific binding, respectively. The sections were then immunostained at 4 °C overnight with the respective primary antibodies: rabbit anti-Iba-1 antibody (1:500, Wako, Osaka, Japan), rabbit anti-MPO antibody (1:100, Abcam, Cambridge, UK) and, rabbit anti-CD3 antibody (1:100, Abcam). After primary antibody incubation, sections were incubated with biotinylated anti-rabbit IgG secondary antibody (1:500, Vector, Burlingame, CA, USA), and then overlaid with avidin–biotin horseradish peroxidase (HRP) complex (Vector). 3,3′-diaminobenzidine solution (DAB, Zsgb-bio, Beijing, China) was used to detect the HRP activity under light microscopy. The number of positive cells around the contusional cortex (see Fig. 1c) was calculated in six random microscopic fields (Olympus, Tokyo, Japan) of each section (three sections per animal) by ImageJ software (Version1.46r, Wayne Rasband, National Institute of Mental Health, USA) and the results were presented as the mean number of positive cells per square millimeter in the tiled images. All counts were performed in a blinded fashion.

To better assess neuronal apoptosis in the peri-contusional cortex, immunofluorescent double staining of terminal deoxynucleotidyl transferase-dUTP nick end labeling (TUNEL) and neuronal nuclei (NeuN) was conducted to determine colocalization of apoptotic cells and neurons. In brief, frozen sections were immunostained with mouse anti-NeuN antibody (1:100, Abcam) at 4 °C overnight and subsequently subjected to TUNEL staining using an in Situ Cell Death Detection kit (Roche, South San Francisco, CA, USA) according to the manufacturer’s suggested protocol. Finally, the sections were covered with 4′,6-diamidino-2-phenylindole (DAPI, Invitrogen). Positive cells around the injured cortex (see Fig. 1c) were calculated per square millimeter from six random microscopic fields of each section (three sections per animal) under a fluorescence microscope (Olympus). All counts were performed in a blinded fashion. The results were presented as the number of apoptotic neurons (NeuN-TUNEL double-stained cells) and the apoptotic ratio of the total neurons (NeuN-TUNEL double stained cells/NeuN-stained cells).

Total RNA of each brain sample was extracted using the TRIzol reagent (Invitrogen) according to the manufacturer’s protocol. The concentration of each RNA sample was quantified using ultraviolet spectrophotometry at 260/280 nm, only samples with an A260/280 > 1.8 were used for further analysis to ensure RNA quality. cDNA was transcribed using a SuperScript® III CellsDirect™ cDNA Synthesis Kit (Invitrogen) from 1.5 μg of RNA. qRT-PCR analysis was performed using an Opticon 2 Real-Time PCR Detection System (Bio-Rad, Hercules, CA, USA) with the SYBR® Green PCR Master Mix (Applied Biosystems, USA). The primers were subjected to 40 cycles of amplification at 95 °C for 10 s and 60 °C for 1 min. GAPDH was used as an internal control and the expression level of each target gene was normalized to that of GAPDH using the 2^−ΔΔCt method (Ct = threshold cycle).

The primer sequences were as follows:

MCP-1: forward 5′-GTGCTGACCCCAAGAGGAA-3′,

reverse 5′-TTGTGGAAAAGGTAGTGGATG C-3′

iNOS: forward 5′-TTGGAGCGAGTTGTGGATTG-3′,

reverse 5′-GTGAGGGCTTGGCTGAGTGA-3′

CD11b: forward 5′-CAGGGCAGGAGTCGTATTG-3′,

reverse 5′-GTCCATCAGCTTCGGTGTTG-3′

Arg-1: forward 5′-TGAACACGGCAGTGGCTTTA-3′,

reverse 5′-GTAGTCAGTCCCTGGCTTATGG-3′

YM1: forward 5′-GCAGAATAATGAGATCACTTACACAC-3′,

reverse 5′-ACGAAGGAATCTGATAACTGACTG-3′

CD206: forward 5′-AAACACAGACTGACCCTTCCC-3′,

reverse 5′- GTTAGTGTACCGCACCCTCC-3′

GAPDH: forward 5′- GGTGAAGGTCGGTGTGAACG-3′,

reverse 5′- CTCGCTCCTGGAAGATGGTG-3′

Total protein concentrations were measured using a BCA Protein Assay Kit (Thermo Fisher Scientific, Carlsbad, CA, USA). The levels of transforming growth factor-β1 (TGF-β1), interleukin-10 (IL-10), interferon-γ (IFN-γ), IL-6, regulated upon activation normal T-cell expressed and secreted (RANTES), and interferon-γ inducible protein-10 (IP-10) were measured using specific ELISA kits (Anoric-Bio, Tianjin, China) according to the manufacturer’s instructions.

All the data were presented as the mean ± standard deviation (SD) and were analyzed using SPSS statistical software (version 22.0, IBM). Non-parametric data from the mNSS test were analyzed using the Kruskal-Wallis H analysis followed by a Mann-Whitney U test. One-way analysis of variance (ANOVA) with repeated measures with Bonferroni post hoc comparisons was used to analyze Rota-rod data. The remaining biochemical data were analyzed using a two-way ANOVA (TBI or sham) × (atorvastatin or saline) with post hoc Bonferroni multiple comparison test. The level of significance was set at a P value of <0.05.

To establish a dose-response for acute atorvastatin treatment post-TBI, we compared a range of dosing concentrations with vehicle saline following TBI. Prior to surgery, there were no differences among all the mice in both tests. In the mNSS test (Fig. 1d), mice in the sham groups showed mild neurological dysfunction at 24 and 72 h after craniotomy, while TBI caused significantly higher mNSS scores compared with the sham groups (p < 0.001). No significant differences were observed among the four TBI groups at 24 h post-TBI (p > 0.05). The mNSS scores decreased with time in all TBI groups because of spontaneous recovery. However, the scores were significantly lower in the 1, 5, and 10 mg/kg/day atorvastatin-treated groups compared with the saline-treated group at 72 h post-TBI (p < 0.01, 0.05, and 0.01, respectively). There was no significant difference among the three different dose groups at all tested time points. In the Rota-rod test (Fig. 1e), the sham groups exhibited the best performance compared with the other four TBI groups (all p < 0.001). At the same time, mice treated with 1, 5, and 10 mg/kg/day atorvastatin showed early improvement on the Rota-rod test, starting at 24 h post-TBI, when compared with that of the saline-treated mice, though no significant differences among the three atorvastatin-treated groups were found. Notably, we also observed that 1 mg/kg/day seemed more effective than higher 5 and 10 mg/kg/day, although the differences were not statistically significant.

The dose-response results demonstrated that no benefit of high doses (5 and 10 mg/kg/day) atorvastatin as compared to low dose (1 mg/kg/day) when the treatment was initiated 1 h after TBI, thus, a dose of 1 mg/kg/day was chosen for all subsequent experiments to minimize the possible side effects.

To examine the effect of atorvastatin treatment on subpopulations of brain-invading leukocytes, quantitative and positioning analyses were assessed by flow cytometry and immunohistochemistry. In the flow cytometry analysis, the gating strategy is shown in Fig. 2a. We found that administration of atorvastatin reduced counts of T cells (CD45⁺CD3⁺) significantly in the brain at 72 h post-TBI when compared with the TBI + saline group (Fig. 2e, p < 0.001), whereas no changes in the invasion of B cells (CD45⁺B220⁺) were observed (Fig. 2f, p > 0.05). Moreover, significantly fewer neutrophils (CD11b⁺CD45^highLy6G⁺, Fig. 2d) and NK cells (CD45⁺NK1.1⁺, Fig. 2e) were detected in brain samples of the TBI + atorvastatin group at 24 and 72 h post-TBI compared with those in the TBI + saline group. Further immunohistochemical study of brain sections for MPO+ neutrophils and CD3+ T cells (Fig. 3) revealed that TBI induced a massive invasion of circulating neutrophils and T cells, and that these invading leukocytes were predominantly located in the contusional core and margin. Atorvastatin-treated TBI mice had significantly fewer neutrophils in the peri-contusional region at 24 h post-injury compared with those in the saline-treated TBI mice (Fig. 3c, p < 0.001). Furthermore, T cell invasion was also attenuated significantly by atorvastatin at 72 h post-injury (Fig. 3d, p < 0.001). In addition, no significant differences were observed between the two sham groups in all parameters. These results suggested that leukocyte infiltration profile was altered by atorvastatin treatment after TBI.

To evaluate the effect of atorvastatin on Tregs after TBI, the levels of CD4 + CD25 + Foxp3+ Tregs from the peripheral spleen and the injured brain were analyzed by flow cytometry. The gating strategy is shown in Fig. 4a. Compared with the sham groups, the TBI + saline group showed a lower percentage of CD25 + Foxp3+ Tregs in the spleen CD4+ T cells at 24 (p < 0.01) and 72 h (p < 0.05) post-injury (Fig. 4c), but a higher level of Tregs in the brain at 72 h (Fig. 4d, p < 0.001), suggesting that TBI per se may upregulate the migration of peripheral Tregs to the injured brain. However, a significantly higher level of Tregs in the spleen and brain was observed in the atorvastatin group compared with that in the saline group at 24 and 72 h post-TBI. As expected, no difference was observed between the two sham groups. These results indicated that atorvastatin could expand CD4+CD25+Foxp3+ Tregs in both the peripheral spleen and the injured brain follow TBI.

To determine the effect of atorvastatin on the expressions of inflammatory-associated mediators in the peri-contusional cortex, an array of inflammatory cytokines were measured using ELISA. The expressions of all detected cytokines were low in the sham groups. Protein levels of the anti-inflammatory cytokines TGF-β1 at 24 (p < 0.05) and 72 h (p < 0.001), IL-10 at 24 (p < 0.05) and 72 h (p < 0.01) post-TBI, increased in the TBI + atorvastatin group compared with those in the TBI + saline group (Fig. 5a, b). In contrast, levels of the pro-inflammatory cytokines IFN-γ at 24 (p < 0.001) and 72 h (p < 0.001), and IL-6 at 24 (p < 0.01) and 72 h (p < 0.001) post-TBI, all decreased significantly in the TBI + atorvastatin group compared with those in the TBI + saline group (Fig. 5c, d). The chemokines RANTES at 24 (p < 0.001) and 72 h (p < 0.05), and IP-10 at 24 (p < 0.05) and 72 h (p < 0.001) post-TBI, were reduced in atorvastatin group compared with those in the saline group (Fig. 5e, f). Notably, the expressions of all these inflammatory mediators in the sham group were not affected by atorvastatin.

To determine whether acute atorvastatin treatment could affect the microglia/macrophage activation, activated microglia or macrophages were detected by flow cytometry and immunohistochemistry. As shown in Fig. 2a, flow cytometry based gating study was performed to discriminate between brain-intrinsic microglia and infiltrating macrophages. At 72 h post-injury, there was a significantly reduced number of CD11b + CD45low microglia in the injured brain of atorvastatin-treated TBI mice when compared to saline-treated TBI mice (Fig. 2b, p < 0.001). Further analysis with a Ly6G antibody demonstrated that atorvastatin diminished mobilization of CD11b+CD45highLy6G- macrophages post-TBI (Fig. 2c, p < 0.001). Moreover, Iba-1 immunohistochemistry was performed to label total activated microglia/macrophages in the injured brain. As shown in Fig. 6a, the Iba-1+ cells in the sham groups were highly ramified, exhibiting small cell bodies and high degree of ramifications. In contrast, TBI provoked a drastic change from the ramified shape to larger cell bodies with fewer ramifications (hypertrophic or amoeboid shape). Compared with the saline-treated mice, the atorvastatin-treated mice showed smaller soma size and higher ramification index, and had significantly fewer Iba-1+ cells in the peri-contusional region (Fig. 6b, p < 0.05) at 72 h post-TBI. These results suggested that atorvastatin treatment not only decreased total activated microglial/macrophage numbers but also induced a morphological shift from the amoeboid shape toward the surveillant ramified morphology.

We further investigate microglial/macrophage activation by examining their polarization status after TBI. As shown in Fig. 7a, CD11b-positive microglial/macrophages were labeled with markers for M1 (CD86) and M2 (CD206) phenotypes. Flow cytometric analysis clearly demonstrated that atorvastatin treatment would reduce pro-inflammatory M1 cells (CD11b⁺CD86⁺, Fig. 7b, p < 0.05) and increase anti-inflammatory M2 cells (CD11b⁺ CD206⁺ Fig. 7c, p < 0.01), thereby augmenting M2/M1 ratio (Fig. 7d, p < 0.01) at 72 h post-TBI compared with that in the TBI + saline group. In addition, gene expression related to the M1/M2 phenotype in the peri-contusional region of brain tissues was analyzed using qRT-PCR assay. Consistent with the flow cytometry results, TBI induced dramatic increases in the expression of all M1/M2 markers compared with those in the sham groups at 72 h post-injury (Fig. 7e, f). With the administration of atorvastatin, the mRNA expression levels of M1-type genes MCP-1, iNOS, and CD11b were decreased significantly, whereas the expressions of all the tested M2-type gene markers, including Arg1, Ym1, and CD206, increased dramatically compared with those in the TBI + saline group. Taken together, these results demonstrated that acute atorvastatin treatment could significantly alter the M1/M2 phenotype balance, via both inhibiting M1 activation and enhancing M2 activation after TBI.

As depicted in Fig. 8a, double staining of TUNEL and NeuN showed that TUNEL-positive apoptotic cells mainly occurred in neurons. There were almost no TUNEL-positive neurons in the right cortex in the sham groups, and the number of apoptotic neurons in the peri-contusional cortex increased significantly in the TBI groups compared with that in the sham groups (p < 0.001). However, compared with the TBI + saline group, treatment with atorvastatin substantially reduced the number of apoptotic neurons at 72 h post-TBI (Fig. 8b, p < 0.01). Furthermore, the percentage of apoptotic neurons in total NeuN-positive neurons was also analyzed. As shown in Fig. 8c, a larger percentage of apoptotic neurons in TBI groups was observed (p < 0.001), whereas atorvastatin promoted neuronal survival by attenuating the percentage of apoptotic neurons (p < 0.001).