Suppression of TRPM7 by carvacrol protects against injured spinal cord by inhibiting blood-spinal cord barrier disruption

When the blood-spinal cord barrier (BSCB) is disrupted after a spinal cord injury (SCI), several pathophysiological cascades occur, including inammation and apoptotic cell death of neurons and oligodendrocytes, resulting in permanent neurological decits. Transient receptor potential melastatin 7 (TRPM7) is involved in the pathological processes in many neuronal diseases, including traumatic brain injury, amyotrophic lateral sclerosis, parkinsonism dementia, and Alzheimer’s disease. Furthermore, carvacrol (CAR), a TRPM7 inhibitor, is known to protect against SCI by reducing oxidative stress and inhibiting the endothelial nitric oxide synthase pathway. However, the functions of TRPM7 in the regulation of BSCB homeostasis after SCI have not been examined. Here, we demonstrated that TRPM7, a calcium-mediated non-selective divalent cation channel, plays a critical role after SCI in rat. Rats were contused at T9 and given CAR (50 mg/kg) via intraperitoneally immediately and 12 hours after SCI, and then given the same dose once a day for 7 days. TRPM7 was found to be up-regulated after SCI in both in vitro and in vivo studies, and it was expressed in blood vessels alongside neurons and oligodendrocytes. Additionally, CAR treatment suppressed BSCB disruption by inhibiting the loss of TJ proteins and preserved TJ integrity. CAR also reduced apoptotic cell death and improved functional recovery after SCI by preventing BSCB disruption caused by blood inltration and inammatory responses. Based on these ndings, we propose that blocking the TRPM7 channel can inhibit the destruction of the BSCB and it is a potential target in therapeutic drug development for use in SCI.


Introduction
Spinal cord injury (SCI) is a devastating condition that results in permanent disability and thereby in uences a profound effect on the quality of life. It causes dysfunction of the limbs and trunk below the injured region and leads to spinal cord edema, apoptotic cell death and permanent disability [1]. Despite years of intensive research, there are no widely accepted therapies to reduce tissue injury and enhance functional recovery after SCI. Initial mechanical damage causes necrosis and blood vessel rupture at the site of the lesion. After that, a series of pathological events occur in response to the primary injury, resulting in a secondary injury that affects both the wound and the surrounding area. Especially, hemorrhage, impaired blood ow, and blood-spinal cord barrier (BSCB) disruption are critical events that cause secondary damage, which results in chronic permanent dysfunction after SCI [1]. When the BSCB is disrupted, blood cells like neutrophils and macrophages in ltrate the spinal tissue and produce in ammatory mediators like pro-in ammatory cytokines, which contribute to secondary injuries [2,3].
Therefore, the development of therapeutic agents inhibiting BSCB disruption would be useful for the restriction of secondary injury followed functional recovery after SCI.
Extensive ion imbalances, including calcium overload in the spinal cord after a traumatic injury, causes irreversible damage to various cellular functions such as protein synthesis, and mitochondria, cytoskeleton, and cell membrane functions, which is one of the leading causes of cell death [4]. Transient receptor potential (TRP) ion channels, which are non-selective cation channels, are expressed in many cells and tissues, including neurons and endothelial cells of the nervous system and have been shown to play important roles in diverse cellular processes [5,6]. The TRP melastatin 7 (TRPM7) channel, in particular, is a metal ion-permeable, non-selective cation channel that is expressed in almost all tissues [7] and regulates divalent cation (Mg 2+ , Ca 2+ , and Zn 2+ ) homeostasis [8,9], cell survival, proliferation, cell adhesion, and Ca 2+ -mediated neurotransmitter release [10].
TRPM7 has been shown to play an important role under pathological conditions [11], although its function has been mainly researched in neurons. Previous studies indicated that TRPM7 mediates the death of anoxic neurons by regulating Ca 2+ in ux during cerebral ischemia and prolonged oxygenglucose deprivation [12][13][14]. The suppression of TRPM7 after brain ischemia also facilitated neuron survival and preserved the morphology and function of neurons in hippocampal CA1 [15]. Additionally, it was reported that TRPM7 is involved in the pathologic processes of some neurodegenerative diseases including western paci c amyotrophic lateral sclerosis, parkinsonism dementia, and Alzheimer's disease [16,17]. Furthermore, carvacrol (CAR), a TRPM7 inhibitor, is known to protect against SCI by suppressing oxidative damage and the endothelial nitric oxide synthase pathway [18]. However, the functions of TRPM7 in the regulation of BSCB homeostasis after SCI have not been examined.
In the present study, we examined the role of TRPM7 by investigating the effect of CAR on BSCB disruption after SCI in rats. Our study showed that the expression of TRPM7 was up-regulated after SCI, and its expression was observed in blood vessels including neurons and oligodendrocytes. Additionally, treatment with CAR, a TRPM7 inhibitor, reduced apoptotic cell death and improved functional recovery by preventing BSCB disruption followed blood in ltration and in ammatory responses after SCI.

Materials And Methods
Animal model of SCI Adult male Sprague-Dawley rats (250-300 g, Samtako, Osan, Korea) were used in this study. Before surgery, rats were weighed and anesthetized with 500 mg/kg of chloral hydrate (i.p.). Laminectomy was performed at the thoracic 9-10 (T9-T10) level after shaving the back and neck, exposing the cord beneath without the dura disruption. The exposed spinal cord was subjected to moderate injury (10 g x 25 mm) using a New York University (NYU) impactor as previously described [19]. Rats also underwent laminectomy at T9-T10 level without injury for sham control. All animal experiments were performed in accordance with the Guidelines of Animal Care Committee of the Kyung Hee University (permission number: KHUASP(SE)-17-059).

Endothelial Cell Culture and OGD/Reperfusion
American Type Culture Collection (Manassas, VA) provided a mouse brain endothelial cell line (bEnd.3), which was cultured as previously described [20,21]. bEnd.3 cells was seeded onto 6-well plate with 5 x 10 5 cells in each well and subjected to oxygen glucose deprivation (OGD) in a humidi ed anaerobic chamber (APM-30D, Astec, Fukuoka, Japan) as previously described (Park et al., 2020). Then, cells were Page 4/26 replaced under normoxic conditions and the media was changed with 25 mM glucose containing DMEM.
Under normoxic condition, control cells were also cultured in DMEM with 25 mM glucose.
In vitro and in vivo drug administration CAR (5-isopropyl-2-methylphenol, Sigma-Aldrich, St. Louis, MO) was dissolved in 0.9% saline. For in vitro study, CAR (500 µM) was treated for 30 min before OGD. For control, 0.9% saline without drugs was used. For in vivo study, CAR (50 mg/kg) dissolved in 0.9% saline was infected intraperitoneally (i.p.) into injured rats immediately and 8 h after injury and then further administered once a day for 7 d for behavioral tests or for indicated time points for other experiments. 0.9% saline was injected for vehicle group. Sham-operated control rats did not receive any pharmacological treatment.

Measurement of BSCB disruption
As previously described, the permeability of the BSCB was measured using Evans blue dye extravasation [22]. In brief, Evans blue dye (Sigma) was dissolved in 0.9% saline and 5 ml of dye (2 %) was injected via i.p. at 1 d after SCI. The rats were sacri ced three hours later via intra-cardiac perfusion with PBS. The T9 spinal cords segment (1 cm) with lesion epicenter was removed and homogenized in a 50% trichloroacetic acid solution. The level of Evans blue was determined by spectrophotometer (Molecular device, Sunnyvale, CA) at 620 nm (excitation wavelength) and 680 nm (emission wavelength). The quantity of Evans blue Dye in the sample was calculated by plotting a standard curve with known amounts of dye (micrograms/gram of tissue).

Tissue preparation
Injured rats were anesthetized at indicated time points with chloral hydrate and perfused with 0.1 M PBS and then with 4% paraformaldehyde in 0.1 M PBS. The spinal cord (10 mm) was dissected out and postxed by immersing it in the same xative for 5 h and placing it in 30% sucrose in 0.1 M PBS. After embedding the spinal segment for frozen sections, longitudinal or transverse sections were cut on a cryostat at 10 or 20 µm (CM1850; Leica, Germany). For molecular work, the spinal cord segment (1 cm) with the lesion epicenter were also isolated after perfusion with 0.1 M PBS and frozen at -80°C.

RNA isolation and RT-PCR
TRIZOL Reagent (Invitrogen) was used to prepare total RNA, and 1 µg of total RNA was reversetranscribed into rst strand cDNA using MMLV according to the manufacturer's instructions (Invitrogen) as described [22]. The resulting cDNAs were subjected to RT-PCR using a thermal cycler (Takara Bio, Shiga, Japan). The primers used for Trpm7, Il-6, Tnf-α, Cox-2, inos, Mcp-1, Mip-1α, Mip-1β, Gro-α, Mip-2α and Gapdh were synthesized by the Genotech (Daejeon, Korea) and the primer sequences are shown in Table 1. PCR products were separated using agarose gel electrophoresis (1.5 or 3 %) and stained with ethidium bromide. The relative band intensity was measured by the AlphaImager software (Alpha Innotech Corporation) and compared to sham control value. The relative intensity values from three times of experiments were subjected to statistical analysis.

Western blot
Total protein was isolated using a lysis buffer (1% Nonidet P-40, 20 mM Tris, pH 8.0, 137 mM NaCl, 0.5 mM EDTA, 10% glycerol, 10 mM Na 2 P 2 O 7 , 10 mM NaF, 1 µg/ml aprotinin, 10 µg/ml leupeptin, 1 mM vanadate, and 1 mM PMSF) as described previously [22]. SDS-PAGE was used to separate protein samples (30 µg) and transfer them to nitrocellulose membrane (Millipore, Billerica, MA). The membranes were incubated for 1 h at room temperature with blocking solution (Tris-buffered saline containing 0.1 % tween-20 and 5% skim milk or bovine serum albumin), then overnight at 4°C with primary antibodies listed in Table 2. Secondary antibodies conjugated with horseradish peroxidase (Jackson ImmunoResearch, West Grove, PA) were used to detect the primary antibodies and immunoreactive bands were visualized using chemiluminescence (Thermo Scienti c, Rockford, IL). Immunoblot image and the densitometirc values were obtained by using AlphaImager software (Alpha Innotech Corporation, San Leandro, CA). β-tubulin was normalized to internal standard.

Immunohistochemistry
Frozen sections were processed for immunohistochemistry as previously described [23] using antibodies against TRPM7, neuronal-speci c nuclear protein (NeuN), CC1, RECA1, ZO-1, myeloperoxidase (MPO), ED-1, and cleaved caspase-3, which are listed in Table 2. Fluorescein isothiocyanate (FITC)-or cyanin 3conjugated secondary antibodies (Jackson Immunoresearch) was used for double labeling. The manufacturer's protocol was used to label nuclei with DAPI (Molecular Probes). Serial transverse sections were collected every 100 or 200 m section rostral and caudal 3,000 or 4,000 m to the lesion site to quantify MPO or ED-1 intensity and cleaved caspase-3-positive oligodendrocytes (cleaved caspase-3/CC1 double-positive) (total 40 or 60 sections). Digital images of MPO, ED-1-stained tissues, and cleaved caspase-3/CC1 double-positive oligodendrocytes in the WM were obtained and quanti ed using MetaMorph software (Molecular device) as previously described [22]. The value was normalized to the primary antibody omitted control after the threshold value and backgrounds were quanti ed. Serial sections were also stained with Cresyl violet acetate for histological analysis.

Measurement of transendothelial electrical resistance (TEER)
Endothelial permeability was determined by measuring TEER according to the manufacturer's instructions (Millipore) as described [22,23]. Brie y, bEnd.3 cells were seeded (5 x 10 5 cells) and seeded and cultured for 24 h on Transwell inserts (Transwell-COL, Corning) coated with rat tail collagen. Cells were than subjected to OGD/reperfusion injury and TEER was determined using with a Millicell ERS-2 Volt-Ohm Meter (Millipore). The surface area of the Transwell inserts was used to calculate the values in ohms per square centimeter (Ω·cm 2 ).

Cell counting of viable ventral motor neuron (VMN)
Serial transverse spinal cord sections (20 m thickness) were collected every 500 m section rostral and caudal 8 mm to the lesion site and stained with Cresyl violet acetate to count the number of viable motor neurons (VMN). VMN located in the lower ventral horn and larger than half the size of the sampling square (20 x 20 µm) was manually counted and analyzed using MetaMorph software (Molecular Device) as described [22]. counted and quanti ed as described in the previous report [22,24]. All TUNEL analyses were performed by investigators who were unaware of the experimental conditions.

Behavioral tests
Behavioral tests including Basso-beattie-Bresnahan (BBB), inclined plane test, grid walk, and foot print were performed according to previously described [25][26][27][28]. Basso-Beattie-Bresnahan (BBB) locomotion scale, which is a 22-point scale (scores 0-21) was used to evaluate hindlimb locomotor function. For the inclined plane test, rats were placed on the testing apparatus in one of two positions (right side or left side up), and the maximum angle maintained without falling for 5 s was recorded and averaged. Rats were also tested on a horizontal grid to assess their ability to precisely place the hindlimb, and the number of footfalls (mistakes) in foot placement were counted. For the footprint analysis, both forepaws and hindpaws were dipped in nontoxic red and blue dye before walking across a narrow box (1 m long and 7 cm wide). The footprints were then scanned, and resulting digitized images were examined.
Behavioral analyses were carried out by trained investigators who were unaware of the experimental conditions.

Histological analysis of myelin and axon
Rats were perfused after behavioral tests (35 days after injury) for histological analyses, and frozen sections were prepared as described above. For axon density quanti cation, serial transverse sections were stained with an antibody speci c for 200 kDa neuro lament protein (NF200, 1:4,000; Sigma) and axon density within preselected elds (40 × 40 µm, 1,600 µm 2 ) at speci c sites within the ventral and dorsolateral funiculi was determined as previously described [25]. Some sections were stained with 5hydroxytryptamine (5-HT, 1:5,000; Diasorin, Stillwater, MN) and the ABC method was used to detect labeled cells using a Vectastain kit (Vector Laboratories). Selected slides were stained with 0.1 % Luxol fast blue (Solvent Blue 38; Sigma) in acidi ed 95 % ethanol and incubated overnight at 60°C for myelin staining. The differentiation step was then carried out with 0.05 % lithium carbonate, as previously described [29]. MetaMorph software (Molecular Device) was used to create digital images of Luxol fast blue-stained tissues.

Measurement of lesion volume
As previously described, serial longitudinal sections from rats tested for behavioral analyses were used to measure lesion volume [30]. Every 50 µm section was stained with Cresyl violet acetate and the lesion volume was calculated by measuring the area of cavitation at the injury epicenter with a low-power (1.25 x) objective and then using a MetaMorph software (Molecular Device).

Statistical analysis
The data were presented as the mean, standard deviation, or standard error of the mean (SEM). To determine the statistical signi cance of the difference between the vehicle-and CAR-treated groups, the unpaired Student's t test was used. For immunohistochemical, and molecular analyses comparisons were based on a one-way ANOVA. BBB locomotor scale and inclined plane test were analyzed using repeated measurement ANOVA (time vs treatment) with post-hoc Tukey's multiple comparison. SPSS 15.0 (SPSS Science, Chicago, IL) was used for statistical analysis. P < 0.05 was considered to be statistically signi cance.

Results
The expression of TRPM7 is up-regulated after SCI In the event of cerebral ischemia and hypoxia, ionic imbalances, including calcium overload inside the neuronal cells, are the established cellular and molecular mechanisms for ischemia and hypoxic neuronal cell death followed brain damage [31]. An in vivo study also showed that the expression of TRPM7 was increased in hippocampal neurons through PI-3K signal pathway after middle cerebral artery occlusion model (MCAO) [32].
To test the hypothesis that TRPM7 channels may contribute to biological events following SCI, we rst examined the pro le of TRPM7 expression in injured rat spinal cords. TRPM7 expression was markedly up-regulated after SCI ( Fig. 1A and B). RT-PCR and Western blot analysis revealed that TPRM7 expression was increased and peaked at 7 d after SCI. TRPM7 was also mainly localized in neurons in the GM and oligodendrocytes in the WM, in the uninjured rat spinal cord (Fig. 1C, Sham). By doubleimmuno uorescence analysis, TRPM7 was colocalized in neurons (NeuN) and oligodendrocytes (CC1), respectively (Fig. 1D, Sham). Interestingly, TRPM7 expression was increased in the blood vessels at or near the lesion site in the injured spinal cord (Fig. 1C, 7d). Consistent with these results, this up-regulation of TRPM7 expression was not observed in the blood vessels of uninjured spinal cords (Fig. 1D, Sham) and majority of TRPM7-positive blood vessels in the injured spinal cord were also positive for the endothelial cell marker RECA1 (Fig. 1D, 7d). These results indicate that TRPM7 is up-regulated in the blood vessels after SCI.

CAR inhibits BSCB disruption and preserves tight junction (TJ) integrity after injury
It is well known that SCI causes in the breakdown of the BSCB [33], and blocking BSCB disruption enhances functional recovery by alleviating the in ltration of blood cells including neutrophils and macrophages followed in ammatory responses [21]. To determine whether ion in ux through TRPM7 affects the BSCB disruption, we evaluated the effect of CAR, a TRPM7 inhibitor, on BSCB disruption at 1 d after injury by Evans blue assay. Compared with sham group, the amount of Evans blue dye extravasation was markedly increased at 1d after SCI, implying BSCB leakage ( Fig. 2A). Furthermore, CAR administration signi cantly reduced the amount of Evans blue dye extravasation at 1 d after SCI as compared with the vehicle group (Veh 72.69 ± 5.05 vs. CAR 15.61 ± 2.51 µg/g tissue, p < 0.05). Next, we examined the effect of CAR on the TJ protein levels of ZO-1 and occludin at 4 h and 5 d after injury. As shown in Fig. 2B, CAR signi cantly attenuated the decrease in TJ protein levels after injury compared with the vehicle control (ZO-1, Veh 0.68 ± 0.04 vs. CAR 0.94 ± 0.04; occludin, Veh 0.62 ± 0.03 vs. CAR 0.94 ± 0.07, p < 0.05). In addition, double immuno uorescence staining for ZO-1 and RECA-1 showed that the fragmentation of capillary blood vessel was increased after SCI and ZO-1 immunoreactivity was decreased upon injury, compared with the sham control, whereas CAR treatment attenuated the fragmentation of the capillary blood vessels and ZO-1 loss (Fig. 2C), indicating that TRPM7 affected TJ integrity and BSCB disruption upon SCI.
TRPM7 is also up-regulated and regulates the integrity of TJ in OGD-induced bEnd.3 cells To understand more about the molecular mechanism underlying TRPM7-mediated regulation in injured blood vessels, an in vitro OGD/reperfusion model with a bEnd.3 mouse brain microvessel endothelial cell line was used. Under the optimized condition of OGD/reperfusion injury to induce the maximal expression of the TRPM7 gene in bEnd.3 cells (data not shown), the expression of the TRPM7 mRNA and protein was increased in the bEnd.3 cells subjected to 6 h of OGD treatment followed by 0.5 h, 1 h, and 3 h of reperfusion ( Fig. 3A and B). By Western blot analysis with anti-ZO-1 and occludin antibodies, the expression of the TJ proteins, ZO-1 and occludin, was decreased in the OGD/reperfusion injury-induced bEnd.3 cells (Fig. 3C, +OGD). However, CAR substantially attenuated the decrease in ZO-1 and occludin expression in the bEnd.3 cells after OGD/reperfusion injury (Fig. 3C, +OGD/CAR) (ZO-1, +OGD 0.6 ± 0.03 vs. +OGD/CAR 1.0 ± 0.04; occludin, +OGD 0.7 ± 0.07 vs. +OGD/CAR 0.9 ± 0.04, p < 0.05). Consistently, trans-endothelial electrical resistance (TEER) was decreased in the OGD/reperfusion injury-induced bEnd.3 cells compared with the untreated control (Fig. 3D, +OGD). In parallel with these results, the decrease of TEER by OGD/reperfusion injury was signi cantly inhibited by CAR treatment (Fig. 3D, +OGD/CAR) (+ OGD 35 ± 6 vs. +OGD/CAR 72 ± 7, p < 0.05). These results indicate that TRPM7 affects TJ integrity by increasing the degradation of TJ molecules.
CAR inhibits the in ltration of neutrophils and macrophages and the expression of in ammatory factors after SCI BSCB disruption and blood cell in ltration after SCI are known to mediate in ammatory responses, thereby contributing to the secondary injury cascade by releasing in ammatory mediators such as Il-6, Tnf-α, Cox-2, and iNos [20,34,35]. Furthermore, the early increase in chemokines such as Mcp-1, Mip-1β, Gro-α, and Mip-2α following SCI is considered to cause neutrophil and macrophage in ltration, thus facilitating in ammatory responses [36-39]. Since CAR prevented BSCB disruption following, the effect of CAR treatment on blood cell in ltration was examined by immuno uorescence staining and Western blot with neutrophil and macrophage cell markers, MPO and ED-1 antibodies. Immuno uorescence staining showed that numerous MPO-positive cells (after 1 d) and ED-1-positive cells (after 5 d) were observed in the dorsal column of the injured spinal cord. However, CAR treatment attenuated the in ltration of these cells compared with the in ltration into the vehicle control ( Fig. 4A and B).
Additionally, relative uorescence intensity analysis revealed that CAR treatment signi cantly reduced the in ltration of blood cells compared with that of the vehicle control (MPO, Veh 1.0 ± 0.07 vs. CAR 0.6 ± 0.05; ED-1, Veh 1.0 ± 0.02 vs. CAR 0.5 ± 0.08, p < 0.05). By Western blot and quantitative analyses, the level of ED-1 was markedly increased after SCI and was signi cantly attenuated by CAR treatment (Fig. 4C) (Veh 1.0 ± 0.07 vs. CAR 0.6 ± 0.06, p < 0.05). These ndings suggest that the inhibition of TRPM7 reduced the in ltration of blood cells by preventing BSCB disruption after SCI.
Next, RT-PCR and Western blot tests were used to look at the effect of CAR on the expression of in ammatory mediators and chemokines after SCI. The results revealed that the increases CAR inhibited the increases in Tnf-α, IL-1β (at 2 h), IL-6, Cox-2, and iNos (at 6 h) mRNA levels after SCI ( Fig. 4D and E). Furthermore, CAR suppressed the increases in the mRNA levels of Mcp-1, Mip-1α, Mip-1β, Gro-α (at 2 h), and Mip-2α (at 6 h) following injury ( Fig. 4F and G). CAR also decreased COX-2 and iNOS protein levels at 1 d after injury as compared to vehicle control (Fig. 4H).

CAR inhibits the apoptotic cell death of neurons and oligodendrocytes
TRPM7 is known to play an important role in neuronal cell death in various neurodegenerative disease models, including SCI [10,12,18]. Furthermore, after BSCB damage, in ltrated blood cells such as neutrophils and macrophages are known to produce in ammatory mediators such as cytokines and chemokines, which contribute to cell death [40][41][42]. Thus, based on our results showing that TRPM7 is involved in the BCSB disruption after SCI, we next examined whether CAR inhibits apoptotic cell death by attenuating in ammatory responses following BSCB disruption. As previously mentioned, a massive loss of VMN was observed in the lesion area following injury [30], when compared to the vehicle control, CAR treatment reduced VMN loss both rostral and caudal to the lesion epicenter (Fig. 5A). The cleaved caspase-3-positive cells in the WM 5 days after SCI were CC1-positive oligodendrocytes, according to double immuno uorescence staining. Immuno uorescence with cleaved anti-caspase-3 antibody showed that CAR treatment signi cantly decreased the number of active caspase-3-positive cells in the WM at 5 d after injury as opposed to the vehicle control ( Fig. 5B  Next, by TUNEL staining of spinal tissue at 1 d and 5 d after SCI, TUNEL-positive cells were also observed within the lesion site in the GM at 1 d and outside the lesion area (WM) at 5 d. Consistent with our previous reports [24,43], the majority of TUNEL-positive cells in the GM at 1 d were identi ed as neurons.
TUNEL-positive cells in the WM at 5 d were also observed outside of the lesion area, extending the entire length of the section (20 mm) and were known as oligodendrocytes. As shown in Fig. 5F and G, CAR treatment resulted in a signi cant reduction in the number of TUNEL-positive cells when compared with the vehicle control in the GM at 1 d and the WM at 5 d (1 d, Veh 405 ± 36 vs. CAR 262 ± 4; 5 d, Veh 325 ± 12.2 vs. CAR 185 ± 8.3, p < 0.05). As a result, our ndings show that inhibiting TRPM7 with CAR prevents apoptotic cell death in neurons and oligodendrocytes following injury.

CAR increases functional recovery after SCI
To assess the effect of CAR on functional recovery, CAR (50 mg/kg, i.p) was administered immediately and 8 hours after injury and then once daily for 7 days. Functional recovery was assessed 28 d after injury using BBB scale, inclined plane test, grid, and footprint analysis. As shown in Fig. 6A, CAR signi cantly improved hindlimb locomotor function from 14 d to 35 d after injury compared with the vehicle group (At 28 d, CAR 11.8 ± 1.1 vs. Veh 8.8 ± 0.9, p < 0.05). Furthermore, the ability to control and accurately position the hindlimbs was checked on a horizontal grid at 28 d after injury. From 14 to 28 days after damage, the angle of the incline was signi cantly higher in the CAR-treated rats than in the vehicle group ( Fig. 6B) (At 28 d, CAR 66.3 ± 3.1 vs. Veh 55 ± 2.7 %, p < 0.05). As shown in Fig. 6C, the number of mistakes (footfalls on the grid walk) was signi cantly smaller in the CAR-treated group than in the vehicle group (At 28 d, CAR 43 ± 1.9 vs. Veh 73.6 ± 4.2, p < 0.05). Finally, footprint examination indicated that fairly consistent forelimb-hindlimb coordination was observed in both the vehicle-treated and CAR-treated rats at 35 d after SCI, but CAR-treated group was very little toe dragging, compared to inconsistent dorsal stepping and extensive dragging in the vehicle rats, as revealed by ink streaks extending from both hindlimbs (Fig. 6D).

CAR decreases lesion volume and inhibits the loss of axon and myelin after SCI
We performed a histological study of the spinal cord tissues from the animals used in the behavioral tests to con rm the correlation between the behavioral results and histological results such as axon loss, myelin loss, and lesion volume. To determine whether CAR retains axons after injury, immunostaining with anti-NF200 and anti-5-HT antibodies was used to detect the remaining axons, and the density of the preserved axons was measured as described in the Materials and Methods section. In the sham control, NF200-positive axons were dense and axonal packing was uniform in both the ventral and dorsolateral funiculi ( Fig. 7A and B, Sham). However, the axon density was signi cantly decreased in the injured tissue ( Fig. 7A and B, Veh). The number of NF200-positive axons was signi cantly higher in the CAR-treated group than in the vehicle control group in both the ventral and dorsolateral funiculi ( Fig. 7A and B)  .0 %, P < 0.05). Furthermore, the density of the 5-HT serotonergic axons in the ventral horn was higher in the CAR-treated group than in the vehicle control (Fig. 7C). These ndings imply that CAR treatment reduces axon loss following SCI.
Next, the extent of myelin loss after injury was assessed by Luxol fast blue staining. At 35 d after injury, extensive myelin loss near the lesion area was evident in the vehicle-treated group but not in the sham control (Fig. 7D, Veh); however, CAR treatment attenuated myelin loss (Fig. 7D, CAR). Serial longitudinal sections were stained with Cresyl violet and the lesion volume was assessed to evaluate the tissue loss after SCI. Figure 7E reveals that the overall lesion volume was signi cantly reduced in the CAR-treated group compared to the vehicle-treated group at 35 d after injury (Fig. 7E) (CAR 4.5 ± 0.3 vs. Veh 7.6 ± 0.9 mm 3 , P < 0.05).

Discussion
In this study, we demonstrated that TRPM7, a calcium-mediated non-selective divalent cation channel, plays a critical role in BSCB disruption after SCI. Both in vitro and in vivo studies have shown that the expression of TRPM7 was increased in the endothelial cells consisting of BSCB after SCI. In addition, CAR suppressed BSCB disruption by inhibiting the loss of TJ proteins and preserved the TJ integrity after SCI, suggesting that blocking the TRPM7 channel can inhibit the destruction of the BSCB. CAR also decreased the blood cell in ltration such neutrophils and macrophages following SCI, thereby inhibiting the expression of in ammatory factors such as TNF-α, IL-6, COX-2, iNOS and chemokines such as Mcp-1, Mip-1α, Mip-1β, Gro-α, and Mip-2α, resulting in reduced in ammatory responses. CAR treatment decreased apoptotic cell death in neurons and oligodendrocytes and increased functional recovery after SCI. Thus, our ndings suggest that TRPM7 is a potential target in therapeutic drug development for use in SCI.
The normal function of the spinal cord is dependent on the microenvironment. The breakdown of the BSCB after SCI induces changes in the spinal microenvironment by facilitating immune cells in ltration into the spinal cord [33] and triggers the post-traumatic in ammatory response, which results in additional spinal cord damage. MMP-9 upregulation is linked to BSCB disruption because it degrades the basal components of BSCB, such as TJ proteins. Furthermore, blood brain barrier (B-BB) disruption was reduced in MMP-9 knockout mice after cerebral ischemia by reducing the protein degradation of ZO-1 relative to wild type mice [44]. Furthermore, our recent studies have shown that inhibiting MMP-9 expression and activation signi cantly reduces BSCB disruption [29,35,45]. Specially, we recently found that Jmjd3, histone H3K27 demethylase, plays a important role as an epigenetic regulator in MMP-9 expression in vascular endothelial cells after SCI [23]. CAR treatment signi cantly prevented BSCB disruption, according to our ndings. TJ proteins were also degraded after SCI, but CAR treatment retained these molecules considerably. However, it is not yet known how the inhibition of TRPM7 prevents BSCB disruption after SCI. The precise mechanism underlying TRPM7-mediated inhibition of BSCB disruption after SCI will be investigated in a future study.
Reactive oxygen species (ROS) have play critical roles in the apoptotic cell death after SCI [30,46]. After injury, The accumulation of free radicals such as superoxide anion (O 2 • − ), hydroxyl radical, and peroxynitrite triggered apoptotic cell death [47]. Furthermore, it was reported that nitric oxide and peroxynitrite play important roles in the B-BB disruption and endothelial cell permeability [48,49]. Via the RhoA, phosphatidylinositol 3 kinase, and protein kinase B (PKB/AKT) signaling pathways, superoxide anion has also been shown to control endothelial TJ proteins and improve B-BB permeability [49]. The report by Jiang et al. [18] showed that carvacrol exhibits a neuroprotective effect after SCI by inhibiting oxidative stress and the eNOS signaling pathway. It was also reported that the application of 2aminoethoxy-diphenyl borate (2-APB), an inhibitor of TRPM7, or knockdown of TPRM7, protects the neurons from H 2 O 2 − mediated injury [50]. Thus, we cannot exclude the possibility that carvacrol has a direct neuroprotective effect or the possibility that it prevents BSCB breakdown via its antioxidant effects, although we focused on the role of TRPM7 in the disruption of BSCB after SCI in this study. Future studies will be performed to elucidate the precise mechanism underlying TRPM7-mediated cell death and BSCB disruption by ROS after SCI.
In ammation plays a critical role in the secondary damage caused by SCI. BSCB disruption after injury leads to the cell in ltration into the spinal cord parenchyma. These in ltrated immune cells produce in ammatory factors, causing tissue injury, triggering cell death, and impairing locomotor behavior after SCI [51][52][53]. Several studies have also shown that inhibiting neutrophils and macrophages in ltration after injury reduces apoptotic cell death and improves functional recovery [54,55]. In this study, we found that the neuroprotective effect of CAR is mediated by inhibiting BSCB disruption and the in ltration of in ammatory cells after SCI, which alleviates apoptotic cell death and improves functional recovery.
However, according to previous studies, TRPM7 is expressed in neurons or oligodendrocytes [10]. Furthermore, our data showed that TRPM7 is expressed in neurons and oligodendrocytes in the uninjured and injured spinal cord. Thus, it is possible that the neuroprotective effect of CAR can be mediated by directly affecting the ion in ux into neurons and oligodendrocytes, and this possibility will be examined in a future study.
In summary, our ndings suggest an important role for TRPM7 and its potential mechanisms in BSCB disruption, which might aid in gaining further understanding of BSCB disruption and cell death after SCI.

Availability of data and material
The data that support the ndings of this study are available from the corresponding author on request.