Thrombin acts as inducer of proinflammatory macrophage migration inhibitory factor in astrocytes following rat spinal cord injury

Adult male Sprague–Dawley (SD) rats, weighing 180–220 g, were provided by the Center of Experimental Animals, Nantong University. All animal experiments were approved by the Animal Care and Use Committee of Nantong University and the Animal Care Ethics Committee of Jiangsu Province. All rats were housed in standard cages (five rats in each cage) in an environment maintained at 22 ± 2℃ on a 12–12 h light–dark cycle and had free access to water and food.

The number of animals subjected to surgery was calculated by six per experimental group in triplicate. The contusion SCI rat model was prepared as the previous description [50]. In a nutshell, all animals were anesthetized with 10% chloral hydrate (300 mg/kg) administered intraperitoneally. The fur around the surgical site was shaved and the skin was disinfected with iodophor. Then the spinous processes of T8–T10 vertebrae were surgically exposed, and a laminectomy was performed at the ninth thoracic vertebral level (T9) with the dura remaining intact. The dorsal spinal processes T7 and T11 were fixed with the clamps of the impactor device (IH-0400 Impactor, Precision Systems and Instrumentation). The impactor rod was positioned centrally at T9 (about 3 mm in length) over the spinal cord midline, and the contusion was applied by receiving a 150-kilodyne contusion injury. The impact rod was removed immediately, and the wound was irrigated. The procedure was visually checked by formation of hematoma and by paralysis of the hindlimbs after animal awakening from the anesthesia. For drug delivery, a total of 4.5 μl of 5 mM PAR1 inhibitor SCH79797 (R&D systems) was slowly injected intrathecally, prior to the incision suture. As the vehicle for SCH79797 contains dimethyl sulfoxide (DMSO), which is not biologically inert, the 10 mM SCH79797 stock solution was prepared by addition of 1 mg of SCH79797 into 0.225 ml of 2.2 mg/ml DMSO dissolved in 0.1 M PBS. A double dilution (work solution: 5 mM SCH79797 in 0.1% DMSO) and a vortex were needed before use within 2 min. The rats were subcutaneously administered with a prophylactic dose of enrofloxacin (Sigma-Aldrich; 1 mg/kg) once daily for seven days to prevent urinary tract infection following surgery. Manual expression of bladders was performed twice a day until the animals recovered the spontaneous voiding.

Astrocytes were prepared from the spinal cord of newborn SD rats, 1–2 days after birth, and the astrocytes were isolated and cultured according to the previously described methods [31]. Briefly, the spinal cords removed from the spinal canal were placed into 0.01 M PBS containing 1% penicillin–streptomycin. The spinal cord capsule was stripped clean under the microscope, followed by mincing with scissors and digestion with 0.25% trypsin for 15 min at 37 ℃. The digestion was terminated by addition of Dulbecco’s Modified Eagle’s Medium (DMEM)—high glucose medium containing 10% fetal bovine serum (FBS), 1% penicillin–streptomycin, and 1% L-glutamine. The suspension was then centrifuged at 1200 rpm for 5 min and the cells were resuspended and seeded onto poly-L-lysine pre-coated culture flask in the presence of 5% CO₂. The medium was changed every 3 days until the whole flask is covered with cells. After 7–9 days, the culture flask was shaken at 250 rpm overnight to remove non-astrocytes. Astrocytic phenotype was evaluated by cell exhibiting a characteristic morphology and positive staining for the astrocytic marker glial fibrillary acid protein (GFAP). Astrocytes with purity more than 95% are acceptable for subsequent experiments.

To examine the effects of thrombin on MIF production from astrocytes, the serum was removed by washing 3 times during 15 min in serum-free DMEM. The cells were then subjected to stimulation with 0–200 nM rat thrombin (Sigma, T5772) in the presence or absence of selective inhibitor hirudin (Abcam), SCH79797 (R&D systems), tcY-NH₂ (TOCRIS), PDTC (Beyotime), SB203580 (TOCRIS), SP600125 (TOCRIS), or PD98059 (TOCRIS) for 24 h prior to assay.

For knockdown of PAR3 expression in the astrocytes, the cells were cultured on the six-well plates for 24 h, followed by transfection of PAR3 siRNA1 (sense strand 5'-CCA ACA UCA UAC UCA UAA U dTdT-3', antisense strand 5'-U ACA CGG AGG GUA UGA AGA dTdT-3'), or scramble siRNA (sense strand 5'-GGC UCU AGA AAA GCC UAU GC dTdT-3', antisense strand 5'-GC AUA GGC UUU UCU AGA GCC dTdT-3') with iMAX transfection reagent (Invitrogen) for 24 h. The astrocytes were subsequently stimulated by 100 nM thrombin for another 24 h before ELISA or Western blot assay.

The primary astrocytes were planted in 96-well plates with 6 × 10³ cells/well and cultured in an incubator at 37 ℃ and 5% CO₂ for 24 h. The cells were subsequently treated with different concentrations of SCH79797, tcY-NH₂, or hirudin for 24 h, followed by addition of 100 μl MTT working solution (MTT:DMEM medium = 1:9) to each well in a 37 ℃ incubator for 4–6 h. Finally, 100 μl of 20% SDS solution was added for 20 h, and the absorbance was measured with a microplate reader at a 570 nm wavelength.

The animals were intraperitoneally anesthetized at desired time points as mentioned above following SCI and drug administration. Then, they were transcardially perfused with 0.01 M PBS (pH 7.4) and 4% paraformaldehyde. The vertebra segments were harvested from six experimental subjects of each time point, post-fixed with 4% paraformaldehyde in PBS at 4 °C overnight. After they were equilibrated in 10%, 20%, and 30% gradient sucrose, about twenty cord sections at 10 μm each were prepared using a cryostat from 0.25 cm length to the lesion epicenter. The sections were blocked with 0.01 M PBS containing 3% BSA, 0.1% Triton X-100, and 10% normal goat serum for 1 h at 37 ℃, and incubated overnight at 4 ℃ with primary antibodies: GFAP (1:400, Sigma); MIF (1:200, Abcam); and PAR1 (1:200, Novusbio)S100β (1:400, Abcam). Thereafter, the sections were rinsed with PBS and incubated with the Cy3-labeled goat anti-rabbit IgG (1:400, Proteintech) or the Alexa Fluor 488-labeled donkey anti-mouse IgG (1:400, Abcam). Sections were observed under a fluorescence microscope (ZAISS, axio image M2).

Total RNA of astrocytes following treatment with 100 nM thrombin for 6 h, 12 h, and 24 h, respectively, was extracted using the mirVana miRNA Isolation Kit (Ambion, Austin, TX) according to the manufacturer’s instructions. They were then selected by RNA Purification Beads (Illumina, San Diego, CA), and undergone library construction and RNA-seq analysis. The library was constructed by using the Illumina TruSeq RNA sample Prep Kit v2 and sequenced by the Illumina HiSeq-2000 for 50 cycles. High-quality reads that passed the Illumina quality filters were kept for the sequence analysis.

Differentially expressed mRNA was designated in a criterion of greater or less than twofold changes in comparison with the control. Function of genes was annotated by Blastx against the NCBI database or the AGRIS database with E value threshold of 10^–5. Gene ontology (GO) classification was obtained by WEGO via GO id annotated by Perl and R program. Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways were assigned to the sequences using KEGG Automatic Annotation Server (KAAS) online. For all heatmaps, genes were clustered by Jensen–Shannon divergence.

A reconstructed gene network was created using the Ingenuity Pathway Analysis (IPA) Software on the basis of differentially expressed genes to investigate their regulatory pathways and cellular functions.

Total RNA was prepared with Trizol (Sigma) from cells treated with PAR3-siRNA (Ribobio) using a LipofectamineTM RNAiMAX transfection reagent (Invitrogen) for 24 h. Fluorescence-tagged control Cy3 was used as a marker for evaluation of transfection efficiency. Alternatively, total RNA was extracted from 1 cm spinal segments of injured site at 0 day, 1 day, 4 days, and 1 week following contusion (n = 6 in each time point). The first-strand cDNA was synthesized using HiScript II Q RT SuperMix for qPCR (+ gDNA wiper) (Vazyme) in a 20 µl reaction system, and diluted at 1:3 before used in assays. The sequence-specific primers were designed and synthesized by Ribobio. Primer pairs for par1, par3, par4, and gapdh are as follows: forward primer 5'-ACT TCA CCT GCG TGG TCA TCT G-3', reverse primer 5'-ATG GCG GAG AAG GCG GAG AA-3'; forward primer 5'-ATT GGT GTA CCA GCG AAC AT-3', reverse primer 5'-CGT TCC CAT TGA GAT GGT AG-3'; forward primer 5'-GTC AAC GCC TCA CCA CCA TAC T-3', reverse primer 5'-GGA GCC AGC CAA TAG GAA GGT C-3'; and for, forward primer 5'-GGG TCC CAG CTT AGG TTC AT-3', reverse primer 5'-GAG GTC AAT GAA GGG GTC GT-3', respectively. Endpoint PCR was performed with the specific primers and KOD-Plus-Neo (Totybo) master mix using a cycling program: 94 °C, 120 s; 98 °C, 10 s; 58 °C, 30 s; 68 °C, 60 s for 30 cycles on a thermocycler. A total of 10 μl of the PCR products was separated by electrophoresis on 1% agarose gel and the images were captured using Gel imaging system. For quantitative real-time PCR (Q-PCR), reactions were performed in a final volume of 10 μl (1 μl cDNA template and 9 μl reaction buffer containing 5 μl of 2 × ChamQ Universal SYBR qPCR Master Mix, 3 μl of RNase-free ddH₂O, and 0.5 μl of anti-sense and sense primers each). Reactions were processed using one initial denaturation cycle at 94 °C for 5 min, followed by 40 cycles of 94 °C for 30 s, 60 °C for 30 s, and 72 °C for 30 s. Fluorescence was recorded during each annealing step. At the end of each PCR run, data were automatically analyzed by the system and amplification plots obtained. The expression levels were normalized to an endogenous gapdh. The relative expression was then processed using the 2^−ΔΔCT method. In addition, a negative control without the first-strand cDNA was also performed.

The hindlimb locomotor function recovery was evaluated using the Basso, Beattie, and Bresnahan (BBB) locomotor scale as previously described [51, 52]. Briefly, after intrathecal injection of 4.5 μl of 5 mM SCH79797 or vehicle at 0, 7, 14, and 21 days, three well-trained investigators blind to the study were invited to observe the behavior of rats for 5 min. The BBB score ranged from 0 (no hindlimb movement) to 21 (normal movements, coordinated gait with parallel paw placement) according to the rating scale. Scores from 0 to 7 indicated the return of isolated movements in the hip, knee, and ankle joints. Scores from 8 to 13 meant the return of paw placement and coordinated movements with the forelimbs. Scores from 14 to 21 indicated the return of toe clearance during stepping, predominant paw position, trunk stability, and tail position. The tests were independently scored by the investigators following assessments.

Statistical analysis used GraphPad Prism 8 software (San Diego, CA, USA). All data are presented as means ± SEM. Comparisons between two groups following of a normal distribution were analyzed by two-tailed unpaired Student’s t test or the Mann–Whitney test when the distribution was not parametric. Differences between multiple groups were analyzed using one­way or two-way analysis of variance (ANOVA), followed by Tukey’s or Sidak's post hoc test. P value < 0.05 was considered statistically significant and was denoted in the figures as P < 0.05.

To explore the potential regulatory role of thrombin on the astrocytic production of MIF, the protein levels of the serine protease at lesion site were determined at 0d, 1d, 4d, and 7d following rat spinal cord contusion. ELISA results demonstrated that thrombin was significantly activated at 1d and 4d by the cord injury, but returned to the control level at 7d (Fig. 1a). Meanwhile, the protein levels of proinflammatory MIF in the injured tissues were inducibly increased at 1d, 4d, and 7d following SCI, as was determined by Western blots (Fig. 1b). It is known that thrombin mediates cell events through activation of PAR1, 3, and/or 4 receptor [47]. Thus, the semi-quantitative PCR was carried out to examine the transcriptional levels of PAR1, 3 and 4 receptors before and after SCI. The results demonstrated that the constitutive expression of PAR1 in the spinal cord was significantly higher than those of PAR3 or PAR4, with a slight upregulation at 4d following SCI (Fig. 1c, d). However, the expression of PAR3 and PAR4 was shown to be moderately induced by cord injury from 1d onwards, although at a relatively lower abundance (Fig. 1e, f). We next sought to observe the colocalization of PAR1 receptor with GFAP-positive astrocytes. Immunostaining displayed that PAR1 was detectable in the astrocytes before and after SCI (Fig. 1g). Further examination of MIF cellular distribution showed that this proinflammatory mediator was dynamically regulated by S100β-positive astrocytes at 0d, 1d, 4d, and 7d following cord contusion (Fig. 2a, b). The data indicate that SCI-induced elevation of thrombin possibly affects MIF production in astrocytes through activation of PAR receptors.

To unveil thrombin-activated inflammatory pathways of astrocytes, the transcriptome sequencing was performed following stimulation of primary astrocytes with 100 nM thrombin for 6 h, 12 h, and 24 h, respectively. The purity of astrocytes isolated from spinal cord reached more than 95% (Fig. 3a). Semi-quantitative PCR determination recapitulated the high abundance of PAR1 expression in the primary cultured cells (Fig. 3b). Analysis of transcriptome profile revealed that a total of 3139, 3965 and 4081 differentially expressed genes (DEGs) were identified at the 3 time points, with defined criteria of P < 0.05 and a greater or less than twofold changes (Fig. 3c). Integration of these DEGs characterized 1707 functional genes that were dynamically regulated by thrombin (Fig. 3d). GO enrichment demonstrated that these integrated DEGs were involved in regulation of astrocytic biological processes, including response to LPS, cytokine-mediated signaling pathway, positive regulation of cytokine production, and cell–cell adhesion (Fig. 3e). KEGG analysis displayed that the functional relevant signal pathways, such as TNF-α signaling pathway, cytokine–cytokine receptor interaction and NFκB signaling pathway, were driven by these DEGs following thrombin stimuli (Fig. 4a). Heatmap and cluster dendrogram illustrated that these DEGs were dynamically regulated by thrombin at different time points (Fig. 4b). These data drawn from bioinformatics reflect that thrombin is able to activate astrocytic inflammation via modulation of multiple intracellular signals. To identify thrombin-induced MIF production in astrocytes, the ingenuity pathway analysis (IPA) was further performed on the DEGs integrated at 6 h, 12 h, and 24 h. A reconstructed gene network revealed that MIF was highlighted as a hub regulator with the highest weight value in response to thrombin stimuli (Fig. 4c). The results indicate that thrombin is a potential regulator of MIF expression in astrocytes.

To validate the stimulatory roles of thrombin on the astrocytic production of MIF, the primary astrocytes cultured in serum-free medium were challenged with 0–200 nM thrombin for 24 h. ELISA and Western blot determination demonstrated that the protein levels of MIF in the cell supernatants and lysates were inducibly elevated by thrombin stimuli in dose dependence (Fig. 5a–c). Treatment of the cells with 0–10 U/ml selective inhibitor hirudin for 24 h in the presence of 100 nM thrombin significantly attenuated the production of MIF from astrocytes (Fig. 5d–f). As was expected, hirudin at 0–10 U/ml made no effects on the viability of the cells (Fig. 5d). To elucidate the exact PAR receptor(s) of thrombin action in regulation of MIF expression, the astrocytes were incubated with inhibitors of PAR1 or PAR4, or knocked down with PAR3 siRNA in the presence of 100 nM thrombin. The results demonstrated that addition of 0–5 μM PAR1 inhibitor SCH79797 or 0–100 μM PAR4 inhibitor tcy-NH₂ to the culture medium for 24 h significantly alleviated the effects of thrombin-induced MIF production from astrocytes (Fig. 6a–c, g–i). MTT assay excluded the cellular toxicity of the two inhibitors (Fig. 6c, i). However, knockdown of PAR3 with efficient siRNA1 did not affect the protein level of MIF in astrocytes stimulated by 100 nM thrombin for 24 h (Fig. 6d–f). Taken together, thrombin is able to facilitate MIF production from astrocytes through activation of PAR1/4 receptor.

Thrombin activates inflammatory astrocytic responses via multiple signal pathways, such as STAT3 and PLCε/NFκB pathways [43, 53]. Also, it has been shown to stimulate the activity of MAPKs [54]. To elucidate the mechanism of thrombin-mediated MIF production in astrocytes, the MAPKs/NFκB pathway was taken into primary consideration, due to its critical role in inflammatory activation of CNS [55]. Exposure of primary astrocytes to 0–200 nM thrombin for 24 h resulted in the dose-dependent increase of phosphorylated ERK, JNK, and p65NFκB protein levels, but not of P38 (Fig. 7a–e). Addition of 0–5 μM PAR1 inhibitor SCH79797 to the culture medium in the presence of 100 nM thrombin for 24 h significantly inhibited the stimulatory effects of thrombin on the phosphorylation of ERK, JNK, and p65NFκB, with the most efficient dose at 5 μM. However, the activation of P38 remained unaffected (Fig. 7f–j). The results indicate that thrombin is able to activate intracellular MAPKs/NFκB pathway of astrocytes.

To further investigate the role of thrombin-mediated activation of MAPKs on the MIF production, the selective inhibitor of ERK (PD98059), JNK (SP600125), or P38 (SB203580) was thus used to incubate astrocytes in the presence of 100 nM thrombin. Results revealed that treatment of astrocytes with 10 μM JNK inhibitor SP600125 for 24 h was able to attenuate thrombin-induced production of MIF (Fig. 8a, b), while ERK inhibitor PD98059 and P38 inhibitor SB203580 did not. Accordingly, the protein level of thrombin-stimulated MIF in astrocytes was also reduced by 0–100 μM NFκB inhibitor PDTC in dose dependence (Fig. 8c, d). The data indicate that thrombin-activated JNK/NFκB pathway contributes to the astrocytic production of proinflammatory MIF.

As SCI-induced production of MIF is involved in activation of innate immunity and worsens the neuropathology [15], the inhibitory effects of PAR1 on the production of MIF at lesion site, as well as on the functional recovery, were therefore evaluated. A total of 4.5 μl of 5 mM PAR1 inhibitor SCH79797 or vehicle was intrathecally injected at the lesion sites of the cords following contusion. Western blots demonstrated that SCI-induced elevation of MIF protein levels was significantly suppressed by the inhibitor at 1d, 4d, and 7d, in comparison with the control (Fig. 9a, b). Immunostaining was performed to detect the expression changes of MIF in the astrocytes before and after treatment of SCH79797, showing that MIF expression in S100β-positive astrocytes was remarkably decreased (Fig. 9c–h). These data indicate that inactivation of PAR1 is efficient in inhibition of thrombin-mediated MIF production from astrocytes following SCI.

Excessive inflammation always results in second tissue damage characterized by expansion of lesion size and severe loss of motor function. To shed light on the effects of SCH79797, which was shown efficient in suppressing proinflammatory MIF, on the functional recovery of rats after SCI, HE staining was performed to observe the damaged cord tissues. Results demonstrated that intrathecal injection of SCH79797 was able to reduce the lesion area of the cord at 21d following spinal cord contusion in comparison with the vehicle (Fig. 10a, b). BBB scores were subsequently assessed during 3 weeks to evaluate hindlimb locomotor function. Behavioral tests showed that inhibition of PAR1 immediately after SCI significantly improved the functional recovery of the rats (Fig. 10c). The data indicate that inhibition of thrombin-mediated production of MIF in astrocytes is able to ameliorate motor function following rat SCI.