The role of S100B/RAGE-enhanced ADAM17 activation in endothelial glycocalyx shedding after traumatic brain injury

The animals were purchased from the Experimental Animal Center of the Southern Medical University in China (Certification number: SYXK (Yue) 2016–0167). This study was approved by the Animal Care and Use Committee of the Southern Medical University, China (Approval number: L2019134). The experimental procedure followed the United States National Institutes of Health Guide for the Care and Use of Laboratory Animals (NIH Publication No 85–23, revised 1985). Two hundred and twenty-eight specific-pathogen-free male Sprague-Dawley rats aged 12 weeks and weighing 250–300 g were housed individually under controlled environment with a 12-h light/dark cycle and unrestricted access to pellet food and water. Rats were randomly assigned to different experimental groups. All surgical interventions were performed under anesthesia with a mixture of 13.3% urethane and 0.5% chloralose (0.65 mL/100 g body weight, intraperitoneally). All efforts were made to reduce the number of animals used and to minimize animal discomfort.

The TBI animal model was produced by lateral fluid percussion with surgical procedures performed as previously described [17]. Briefly, anesthetized rats were placed in a stereotaxic frame. After incision of the scalp, the temporal muscles were separated and a 4.8 mm craniotomy was drilled (2.5 mm lateral to the sagittal sinus and centered between bregma and lambda). A hollow female Luer-Lok fitting was placed directly over the dura and rigidly fixed using dental cement. Before the induction of trauma, the female Luer-Lok was connected to the fluid percussion injury device via a transducer (Biomedical Engineering Facility, Medical College of Virginia, USA). For the infliction of TBI, a metal pendulum was released from a pre-selected height, leading to a rapid injection of normal saline into the closed cranial cavity. The pulse of increased intracranial pressure of 21–23 ms duration was controlled and recorded by an oscilloscope (Agilent 54622D, MEGAZoom, Germany). The severity of injury inflicted on the animal was altered by adjusting the amount of pressure generated by the pendulum. For the present experiment, a severe injury level was induced (3.5 ± 0.2 atm). The sham group animals underwent identical preparatory procedures, including craniotomy, but were not was not inflicted with TBI. Rats were sacrificed with pentobarbital (50 μg/kg i.p.) at 5 min, 30 min, 1 h, 3 h and 6 h. Blood samples were collected by cardiac puncture, then serum was obtained by centrifuging the blood at 3000 r/min. The brain, and lungs were carefully collected at the same time. The general pathological changes of the brain were observed by detection of brain histology and water content at 5 min, 30 min, 1 h, 3 h and 6 h following TBI. The expression of S100B and RAGE in brain tissue, the levels of S100B, sRAGE in serum were measured in sham and TBI group at 5 min, 30 min, 1 h, 3 h and 6 h following TBI.

Rats were randomly assigned to different experimental groups. To clarify the effects and signaling relationship of S100B, RAGE, ADAM17 on the development of TBI and the secondary injury of brain and lungs, rats were randomized into sham group, TBI group, TBI + ONO-2506 (S100B chemical inhibitor) group, TBI + FPS-ZM1 (RAGE chemical inhibitor) group, and TBI + TAPI-1 (TNF-alpha processing inhibitor-1, an ADAM17 inhibitor) group. ONO-2506 [18] (25 mg/kg; GLPBIO GC15105, USA), FPS-ZM1 [19] (5 mg/kg; GLPBIO GC10652, USA), or TAPI-1 [20] (10 mg/kg; GLPBIO GC12344, USA) was, respectively, applied through intraperitoneal injection immediately after the onset of TBI, and the relevant parameters were measured 6 h after the onset of TBI.

The cortex of injured-side brain and the lungs from sham and TBI rats were quickly excised, sliced into transverse or longitudinal sections, and fixed in 10% neutral-buffered formalin. The tissues were then embedded in paraffin blocks, and serial sections were stained with hematoxylin and eosin (H&E) for microscopic evaluation at 50, 100 or 200× magnification. Morphological changes were blindly assessed and graded by two certified pathologists using the injury score developed.

The rats were sacrificed at 5 min, 30 min, 1 h, 3 h and 6 h following TBI, the whole brain and lung were excised and weighed. Then, the tissues were placed in an oven at 80 °C loosely wrapped in filter paper for 72 h to achieve dry weight, and the water content of tissues was calculated with ratio of (wet weight–dry weight)/wet weight to assess brain edema.

Twenty specific-pathogen-free male Sprague-Dawley rat pups aged 1–4 days used for astrocyte isolation were purchased from the Experimental Animal Center of the Southern Medical University in China (Certification number: SYXK (Yue) 2016–0167). This study was approved by the Animal Care and Use Committee of the Southern Medical University, China (Approval number: L2019134). The experimental procedure followed the United States National Institutes of Health Guide for the Care and Use of Laboratory Animals (NIH Publication No 85–23, revised 1985).

The protocol of primary astrocyte culture was similar to that described previously [21]. Cortices from 1 to 4-day-old rat pups were used for astrocyte isolation. After decapitation of the rats and removal of the meninges, the cortices were digested by 0.25% Trypsin, and the cells were treated with DNase I (Sigma) before centrifugation at 1000×g for 3 min. Cells were cultured on poly-l-lysine-coated dishes at a density of 1 × 10⁶ cells per 10-cm dish in minimum essential medium with 10% fatal bovine serum. The medium was changed to new medium on the following day. The medium was changed every 3 days, the density of cells can reach 80–90% at 7 days. Subsequently, the primary astrocytes were collected and identified with glial fibrillary acid protein (GFAP) expression through immunofluorescence to assess nerve cell purity. Next experiment would be carried out, while the astrocyte purity reached 99%.

The protocol of astrocyte stretch injury model was similar to those described previously [22]. Equiaxial stretch (20% strain, 1.0 Hz frequency) was applied to cultured astrocytes for 1 h, by a Flexcell^® FX-5000™ Tension System (Flexcell, USA) with specific designed 6-well plates (BioFLEX^®). The culture media of astrocytes in stretch injury group and control group were collected at this time and will be used as conditioned medium. Again, to clarify the effects and signaling relationship of S100B, RAGE, ADAM17 on stretch injury, ONO-2506 (100 µM) [18], FPS-ZM1 (100 nM) [23] or TAPI-1 (50 µM) [24] were, respectively, applied for 6 h after stretch injury. To further verify the effect of S100B on astrocyte injury, exogenous S100B (1 µM) [8] was directly administrated to normal astrocytes for 6 h. The cell viability and relevant parameters were detected as indicated.

A cell counting kit (CCK-8 kit; Beyotime Biotechnology; C0038; China) was used to detect cell viability and cells were cultured in the relevant condition as above. Cells from different treatment groups were counted, adjusted the concentration to 1 × 10⁵ mL, and seeded in a 96-well plate with 100 μL/well. Cells were seeded in triplicate for each treatment group. The 96-well plate was placed in the incubator (37 °C and 5% CO₂) and cells were cultured till the indicated time. 10 μL CCK-8 solution was added to each well and the cell culture plate was incubated for 1 h. The absorbance at 450 nm was detected using a plate reader. Blank wells (culture media and CCK) and control wells (untreated cells, culture media, and CCK) were also detected.

Twenty specific-pathogen-free male Sprague-Dawley rats aged 12 weeks and weighing 250–300 g used endothelial cell isolation were purchased from the Experimental Animal Center of the Southern Medical University in China (Certification number: SYXK (Yue) 2016–0167). This study was approved by the Animal Care and Use Committee of the Southern Medical University, China (Approval number: L2019134). The experimental procedure followed the United States National Institutes of Health Guide for the Care and Use of Laboratory Animals (NIH Publication No 85–23, revised 1985). The animals were housed individually under controlled environment with a 12-h light/dark cycle and unrestricted access to pellet food and water. Rats were randomly assigned to different experimental groups. All surgical interventions were performed under anesthesia with a mixture of 13.3% urethane and 0.5% chloralose (0.65 mL/100 g body weight, intraperitoneally). All efforts were made to reduce the number of animals used and to minimize animal discomfort.

A TBI-induced endothelial injured model was produced using primary rat aortic endothelial cells (RAECs) stimulated with conditioned medium from the astrocyte stretch injury model. The protocol for the isolation and identification of RAECs were modified from Zhu et al. [25]. The animals were first anesthetized and disinfected for the exposure of the aorta. The aorta was quickly isolated after flushing the blood through the left ventricle with PBS and dissected to remove adipose tissue and small colateral vessels. The aortic tissues were then cut into 2 to 5 mm long rings and each aortic ring was cut open and immediately placed with the lumen side down into a six‐well plate containing up to 50 µl of endothelial cell growth medium with endothelial cell growth factor, fetal bovine serum, penicillin, and streptomycin (Cat #1001; ScienCell, Carlsbad, CA). When the sprouting and cell growth reached 80% confluency, the aortic segments were gently removed, the cells were harvested and the von Willebrand factor (vWF) expression was detected through immunofluorescence to identify endothelial cells and to assess purity.

The above-mentioned conditioned medium from astrocyte stretch injury model was applied to RAECs for 6 h, with supplement of vascular endothelial cell growth factor (Cat #1052; ScienCell, Carlsbad, CA). To further verify the effect of S100B on endothelial cells, exogenous S100B (1 µM) (GuangZhou BlueLife Biotechnology Co., Ltd) was directly used to stimulate normal RAECs. To clarify the effects and the signaling relationship of S100B, RAGE, and ADAM17 on glycocalyx shedding of endothelial cells, ONO-2506 (100 µM), FPS-ZM1 (100 nM) or TAPI-1 (1 µM) were added for 6 h while incubating with conditioned medium. The relevant parameters were detected as indicated. The experimental protocol in endothelial injury model is shown the lower right panel of the flow chart in Fig. 1.

The following primary antibodies were used for western blotting or immunofluorescence: GAPDH (AF7021; Affinity), S100B (ab52642; Abcam), RAGE and sRAGE (ab216329; Abcam), GFAP (ab7260; Abcam), vWF (ab6994, Abcam), syndecan-1 (ab128936; Abcam), ADAM17 (ab39162, Abcam), GM130 (DF7556; Affinity). The following secondary antibodies were used: peroxidase-conjugated goat anti-rabbit, Alexa Fluor 488- and Alexa Fluor 594-conjugated secondary antibody were bought from the Beijing Ray Antibody Biotech Company.

Samples were separated by SDS-PAGE under reducing conditions and transferred onto polyvinylidene difluoride (PVDF) membranes (Immobilon-P, Millipore). Membranes were blocked with 5% BSA in TBS (50 mM Tris, 150 mM NaCl, pH 7.4), and subsequently incubated with primary antibodies in 0.1% Tween-TBS with 1% BSA and washed in 0.1% Tween-TBS. Bound primary antibodies were detected with peroxidase-conjugated goat anti-rabbit antibodies using the ECL detection system Super Signal West Pico (Thermo Fisher Scientific, Germany).

After fixation, dehydration, the frozen sections of SD rat brain and lung tissues were prepared. Cells were grown on a confocaldish and treated as indicated. The tissue sections and cells were fixed with 4% methanol, perforation with 0.1% Triton x-100 and blocked for 1 h in 5% BSA and 0.1% Tween-20 in PBS at room temperature. The primary antibodies were applied to the sample, followed by Alexa Fluor 488- and Alexa Fluor 594-conjugated secondary antibody. The images were obtained by laser scanning confocal microscopy (LSCM; Carl Zeiss Microscopy/Zeiss LSM780; Germany). The immunofluorescence intensities of target proteins in brain tissue or cultured cells from five fields per sample in each group were quantified and reported as relative fluorescence units (RFUs).

The onset of TBI was characterized by brain hyperemia, tissue contusion and swelling in the area of impact, and there was obvious hematocele in the ventral surface of the brain (Fig. 2A). The water content in the injured brain began to increase significantly in a time-dependent manner 30 min after TBI (Fig. 2B). H&E staining showed that the brain tissue in sham group were arranged uniformly and the intercellular space was normally distributed. Nerve cells were accompanied in abundant cytoplasm with deep-dyed big nucleolus. The structure of the brain tissues obtained from injured rats was damaged directly through the mechanical impact reflected by ruptured blood vessels and bleeding, resulting in subsequent diffuse hemorrhagic foci. Nuclei became condensed with deeper staining. With the expansion of the hemorrhagic foci, the nerve cells swelled and the cytoplasm vacuolated. Infiltration of inflammatory cells, or obvious leakage of red blood cells could be observed after experimental TBI and this pathological damage increased significantly with the prolonged observation times (Fig. 2C). There was certain mortality in this severe TBI rat model at different timepoints (5 min, 30 min, 1 h, 3 h, 6 h), respectively (Table 2).

Compared with those in the relevant areas in sham group, the expression of S100B and RAGE in cerebral cortices at the edge of injury was increased significantly after TBI (Fig. 2D). Serum S100B and sRAGE levels also increased gradually and significantly after TBI (Fig. 2E). The level of sRAGE in serum was increased at 1 h after TBI, which was earlier as the increased expression of RAGE in the brain tissue (Fig. 2E). The results from immunofluorescence under confocal microscope of brain cerebral cortices were consistent with those from western blot. Precisely, the enhanced staining of cytosolic S100B was co-localized with GFAP-positive (GFAP⁺) cells and membrane RAGE was also increased in the area of GRAP⁺ cells at 6 h after TBI (Fig. 2F).

Cultured primary rat astrocytes were dispersed evenly with flat-spindle-shaped in “slab stone”-like arrangement under light microscope and were identified as GFAP⁺ cells (Fig. 3A). The cells presented with morphology change, and poor cell refraction after mechanically stimulated with stretching (Fig. 3B). Compared with control group, cell viability decreased in a time-dependent manner (Fig. 3C). Stretch injury induced the elevation of S100B in astrocytes and medium. Both stretch injury and exogenous S100B could enhanced the expression of RAGE in astrocytes and sRAGE in medium (Fig. 3D). These data in Figs. 2 and 3 suggest that TBI triggered the up-regulation of S100B/RAGE signal in the pathogenesis of primary and secondary injury.

The administration of ONO-2506 in rat attenuated TBI-induced increase of RAGE expression in cerebral cortices and sRAGE secretion to the serum (Fig. 4A, E). Consistently, the administration of S100B inhibitor ONO-2506 abolished the stretch injury-induced increase of RAGE expression and sRAGE secretion in astrocytes (Fig. 4C).

In addition, it was found that the usage of RAGE inhibitor FPS-ZM1 could reduce the expression of S100B, and subsequently attenuated RAGE expression and sRAGE secretion in TBI rat model (Fig. 4B, E) and astrocyte stretch injury (Fig. 4C). In addition, the administration of RAGE inhibitor FPS-ZM1 could reduce the exogenous S100B-induced increase of S100B expression and secretion in astrocytes (Additional file 1: Fig. S1). The data in Fig. 4 indicate that the release of S100B from astrocytes could promote RAGE expression and sRAGE release after TBI. The enhancement of RAGE, in turn, may have induced further up-regulation of S100B.

The results further demonstrated that the amount of syndecan-1, the glycocalyx damage markers, was decreased in brain tissue from peri-injury cortex, while serum level of syndecan-1 was increased (Fig. 5A, B). Our data revealed similar changes of syndecan-1 in a time-dependent pattern (Additional file 1: Fig. S2).

Consistently, the results from TBI rats treated with ONO-2506 or FPS-ZM1 showed a significant improvement of EG integrity with syndecan-1 more in brain tissue and less in the serum (Fig. 5A, B). The localization of syndecan-1 on the intima of micro vessels was mostly preserved by the application of ONO-2506 or FPS-ZM1 (Fig. 5F).

The results from RAECs stimulated with conditioned medium of stretch injury astrocytes or exogenous S100B revealed that the protein level of syndecan-1 was significantly lowered in the cells and elevated in cultured medium (Fig. 5C). The changes of syndecan-1 in conditioned medium-treated group were attenuated by S100B inhibitor ONO-2506 (Fig. 5D) or RAGE inhibitor FPS-ZM1, respectively (Fig. 5E). The immunostaining of syndecan-1 in RAECs confirmed these results (Fig. 5G). The data from Fig. 5 prove that the activation of S100B/RAGE signaling mediates EG damage after TBI.

The results from TBI rats revealed that the protein levels of pADAM17 (the immature form) and mADAM17 (the mature form) from peri-injury cortex were both higher than those in sham group, along with the increase of syndecan-1 in serum, while the application of ADAM17 inhibitor TAPI-1, a specific hydroxamate inhibitor of metalloprotease disintegrins, effectively blocked this TBI-induced syndecan-1 shedding (Fig. 6A). Similar results were obtained under confocal immunofluorescent, and TAPI-1 specifically preserved the localization of syndecan-1 on the intima of microvessels (Fig. 6C).

The results from RAECs cultured in conditioned medium consistently demonstrated that the protein levels of pADAM17 (the immature form) and mADAM17 (the mature form) were both higher than those in control or sham group, accompanying with the increased release of syndecan-1 in the medium, thus indicating the EG shedding. This induced syndecan-1 shedding was effectively blocked by ADAM17 inhibitor, TAPI-1, with no significant change in pADAM17 and mADAM17 protein levels (Fig. 6B). The immunostaining of syndecan-1 in RAECs confirmed these results (Fig. 6D). These data in Fig. 6 indicate that enhanced expression and the activation of ADAM17 might contribute to the shedding of syndecan-1 and TAPI-1 might be effective on the preservation of glycocalyx after TBI.

While the above-mentioned results have proved that the activation of S100B/RAGE signaling mediates EG damage, it is worthwhile to elucidate whether S100B/RAGE signal could have been involved in the regulation of ADAM17 after TBI. The protein levels of pADAM17 and mADAM17 were both significantly elevated in brain tissue and mADAM17 level was also increased in serum from TBI rats, indicating the secretion of mADAM17 from cells. The administration of S100B or RAGE inhibitor abolished TBI-induced up-regulation of pADAM17 and mADAM17 in brain tissue, and attenuated the secretion of mADAM17 into serum (Fig. 7A, B). The immunostaining of injury-adjacent brain tissue revealed a significant increased co-localization of ADAM17 with endothelial marker vWF and the inhibition of S100B or RAGE attenuated this increase (Fig. 7F).

The results from RAECs also revealed that the protein levels of pADAM17 and mADAM17 were significantly elevated in RAECs cultured with exogenous S100B. It is interesting to note that mADAM17 was also elevated in RAEC medium, indicting the secretion of mADAM17 from endothelial cells (Fig. 7C). Compared to conditioned medium-treated alone, the addition of S100B inhibitor ONO-2506 (Fig. 7D) or RAGE inhibitor FPS-ZM1 (Fig. 7E) exhibited a significant reduction in pADAM17 and mADAM17 expression and secretion of mADAM17. These data from Fig. 7 imply that activating S100B/RAGE signaling may promote ADAM17 expression and activation.