Troxerutin and Cerebroprotein Hydrolysate Injection Protects Neurovascular Units from Oxygen-Glucose Deprivation and Reoxygenation-Induced Injury In Vitro

Cerebral ischemia/reperfusion (I/R) injury involves complex events of cellular and molecular processes. Previous studies suggest that a neurovascular unit (NVU) acts as an intricate network to maintain the neuronal homeostatic microenvironment. The present study established an NVU model for oxygen-glucose deprivation and reoxygenation (OGD/R) damage, trying to target the major components of the NVU using a coculture of rat neurons, astrocytes, and rat brain microvascular endothelial cells (rBMECs) to investigate the therapeutic effects of troxerutin and cerebroprotein hydrolysate injections (TCHis). The study observed that OGD/R downregulated the expressions of GAP-43, Claudin-5, and AQP-4 obviously detected by Western blotting and immunocytochemical analysis, respectively, while TCHi ameliorated the effect of OGD/R significantly. Meanwhile, TCHi alleviated the abnormalities of ultrastructure of neurons and rBMECs induced by OGD/R. Furthermore, both levels of inflammatory cytokines (IL-1β, IL-6, and TNF-α) and cell adhesion molecules (VCAM-1 and ICAM-1) detected by ELISA in NVU supernatant were found elevated significantly through OGD/R, but TCHi ameliorated the trend. In addition, TCHi also mitigated the increase of proapoptotic factors (Bax, p53, and caspase-3) induced by OGD/R in NVU model statistically. All these findings demonstrated that TCHis played a protective role, which was reflected in anti-inflammation, antiapoptosis, and blood–brain barrier maintenance. The results of the study concluded that the NVU is an ideal target and TCHi acts as a neuroprotective agent against cerebral I/R injuries.


Introduction
Cerebral ischemia/reperfusion (I/R) injury is a complex pathological process in the nervous system, resulting in high disability and mortality worldwide, with significant clinical and socioeconomic impacts [1]. The complex pathobiological mechanisms of this medical problem include inflammation, apoptosis, oxidative damage, and ionic imbalances [2]. Although reperfusion after ischemia is essential for cell survival, it may have numerous negative consequences such as microvascular damage, cell dysfunction, and cell death. It can attract leukocytes and cause the release of several proinflammatory mediators and the induction of microglia and macrophages [3]. Excessive neuroinflammation can increase 2 Evidence-Based Complementary and Alternative Medicine [7,8]. Cerebroprotein hydrolysate (with abundant bioactive peptides) was found to facilitate the distribution of troxerutin and had a positive synergistic effect with troxerutin against acute ischemic stroke [9]. The present study aimed to investigate the protective effects of troxerutin and cerebroprotein hydrolysate injections (TCHis) on oxygen-glucose deprivation and reoxygenation-(OGD/R-) inducing NVU dysfunction and the possible mechanism.

Animals.
Adult Wistar rats (3 months old) were purchased from Peking Vital River Laboratory Animal Ltd. Three female rats were mated with one male rat in each cage, and the pregnant females were kept individually. The rat pups were used for further experiments. All experiments were performed in accordance with China's Guidelines for Care and Use of Laboratory Animals.

Primary Cell Cultures.
Primary cells were extracted from the rat pups and routinely cultured in conditioned incubators (37 ∘ C/5% CO 2 ). The isolation procedure was performed according to the methods used by Xue et al. (2013) [10] and Wang et al. (2015) [11].
Primary cortical neurons (N) were prepared from Wistar newborn rats (less than 24 h). In brief, the cerebral cortex was digested with 0.125% trypsin for 10 min at 37 ∘ C and the cell suspension was passed through a 75 m pore filter. Cells were harvested and seeded on poly-D-lysine (Sigma Aldrich, MO, USA) precoated plates in Neurobasal Medium (Invitrogen, CA, US) containing 2% B27 supplement (Invitrogen), 1% penicillin-streptomycin, and 2 mM L-glutamine. Experiments were performed for 8 days in vitro.
Primary astrocytes (A) were extracted from 1-to 2-dayold rat pups, as described previously with a few modifications (Saini MG et al., 2011). In brief, the cerebral cortex was digested with 0.125% trypsin for 10 min at 37 ∘ C, and the cell suspension was then passed through a 75 m pore filter. Cells were seeded in the DMEM/F12 medium containing 10% fetal bovine serum (FBS) (NQBB, Australia) and 1% penicillin-streptomycin. After 7-10 days, the cultures were shaken at 37 ∘ C at a speed of 260 rpm for 16 h to remove contaminating microglia and oligodendrocytes. The third passage of astrocytes was used for the following study.
Primary brain microvascular endothelial cell (rBMEC) cultures were established from 7-day-old rat pup brain tissues (B), which were extracted and homogenized with type II collagenase/DNaseI (Sigma) for 1 h. Microvessels were separated after density centrifugation (spun at 1000 g/min for 20 min) in 20% bovine serum albumin at 4 ∘ C. The microvessels were then digested using a collagenase/dispase solution (Roche Applied Science, Mannheim, Germany) containing DNaseI for 1 h and suspended in a DMEM high-glucose medium containing 20% FBS, 10 ng/mL basic fibroblast growth factor, 30 U/mL heparin, 2 mM glutamine, and 1% penicillin-streptomycin. Cells were then seeded into gelatin (1%) coated flasks, and the third passage of rBMECs was used in this study.

Establishment of NVU In
Vitro. The NVU model was established according to the previous report using purified normal morphological cells ( Figure 1) [12]. Briefly, after the neurons had grown in a six-well culture plate with a density of 0.5 × 10 5 cells/cm 2 for 2 days, the purified astrocytes with 1.5 × 10 5 cells/cm 2 were seeded on the outer side of the insert membrane, which faced the bottom of the well. After 4 h for astrocyte adhering, the insert was placed into the well with neurons. Two days later, rBMECs (1.0 × 10 5 cells/cm 2 ) were seeded in the inner side of the insert membrane. After being cocultured for 3-5 days, the NVU model was prepared for the following experiments (Figures 1 and 2).

Four-Hour Leakage
Detection. Blood-brain barrier (BBB) permeability was evaluated by performing a 4-hour leakage experiment. After 3 days of coculture, the upper inserts were filled with the medium, while the level of the medium in the plates was maintained 0.5 cm lower than the level of the medium in the upper inserts. Inserts with no cells were used as a control. After 4 h, changes in the level of the medium in the top inserts were observed ( Figure 2).   [14] mentioned in their study. The cocultures were postfixed in 1% osmium tetroxide (Electron Microscopy Sciences, Inc., PA, USA) in 0.1 M PB for 1 h at room temperature. Following osmication, they were dehydrated in a graded series of acetone dilutions (30%, 50%, 70%, and 95% acetone in water), allowing 10 min for each change. Three 10 min changes in 100% acetone were made, and the cocultures were transferred to a 50 : 50 mix of acetone: LX112 epoxy resin embedding mix (Ladd Research Industries, Burlington, VT). Subsequently, the cocultures were infiltrated with this mix for 1 h under vacuum and were transferred to a 100% LX112 embedding mix and infiltrated for 1 h on a rotator. Two more 1 h infiltration steps were performed with fresh changes of the embedding mixture. Cocultures were further infiltrated in a fresh embedding medium at 4 ∘ C overnight. The following    , and DAPI is labeled in blue). * < 0.05, * * < 0.01, and * * * < 0.001 relative to Model; ### < 0.01 relative to CK.

Statistical Analysis.
All the results were repeated at least three times. All pictures were analyzed by Adobe Photoshop software. Data were statistically analyzed by analysis of variance (using IBM SPSS 17.0). Results were expressed as mean ± SD. < 0.05 was regarded to be statistically significant.

Result
Morphology and specific identification for three types of cells in the coculture system: as shown in Figure 2, pictures of the same field were obtained in visible light under the fluorescent inverted microscope. Neuronal cells were cultured at the bottom of the Transwell filter (Figure 2(a)), rBMECs (Figure 2(b)) were seeded on the upper side of the Transwell filter of the inserts, and astrocytes were seeded on the opposite side of the Transwell filter of the inserts (Figure 2(c)). The schematic drawing of the triple cell coculture system is shown in Figure 1. The transendothelial electrical resistance (TEER) of different models indicated that the coculture system had an acceptable BBB function (Figure 2(d)). TEM findings demonstrated that BBB appeared normal in rBMEC; meanwhile, tight junctions and desmosomes were close and adjacent (Figure 3).

Effects of TCHi on Cell Survival in NVU Cells after OGD/R.
As shown in Figure 4, neurons, astrocytes, and rBMECs had typical damage manifestations after OGD/R. However, with the treatment of TCHi at doses of 10 M, 100 M, and 1000 M, these manifestations were weakened. Western blotting was analyzed to confirm the effect of TCHi, and the findings were similar to those of immunocytochemical analysis (Figures 4 and 5).

Effects of TCHi on Antiapoptosis in the NVU Model after
OGD/R. Whether TCHi had effects on apoptosis was further checked. Proapoptotic factors, Bax, p53, and caspase-3, were detected using immunocytochemical analysis. Figure 6 showed that TCHi at concentrations of 10 M, 100 M, and 1000 M suppressed all these factors.  , and (c) reflect neurons, rBMECs, and astrocytes, respectively (Bax is labeled in orange, p53 is labeled in red, caspase-3 is labeled in green, and DAPI is labeled in blue).

Effects of TCHi on the Ultrastructure of NVU Cells after
OGD/R. The ultrastructures of neurons and rBMECs were observed by TEM to confirm the effectiveness of TCHi. Figure 7 indicated that both neurons and rBMECs showed apoptotic signs after OGD/R. Cells exhibited shrinkage of shape, irregular nuclei, diffused distribution of heterochromatin, autophagosome appearance, and so on. At concentrations of 1000 M, TCHi reversed all these signs.

Discussion
Our study demonstrated that an injection of TCH could ameliorate OGD/R, inducing NVU dysfunction. The mechanism of TCH treatment was to suppress inflammation and apoptosis. All these findings in the NVU model implied the underlying therapeutic effect of TCHi against cerebral ischemic strokes, which might explain positive clinical findings for TCHi [9]. Neuroinflammation is a complex inflammatory process in the central nervous system, which is thought to play an important defensive role against various pathogens [15]. However, an aberrant inflammatory response is known to be a type of reperfusion injury after a stroke [16]. Several cytokines and chemokines are released after an ischemic brain injury. The most extensively studied of the proinflammatory cytokines include IL-1 , IL-6, and TNF- [16]. This study measured the concentrations of IL-1 , IL-6, and TNFin the OGD/R model and found that TCHi has an inhibitory effect on all these cytokines.
An inflammatory process consists of both the activation of resident cells of the central nervous system and the infiltration of peripheral leukocytes into the ischemic   brain tissue [17]. Cell adhesion molecules are involved in the trafficking and recruitment of leukocytes to activate ischemic endothelia in strokes, which could worsen ischemic brain injuries [18]. ICAM-1 and VCAM-1 are reported to be upregulated by the proinflammatory cytokines TNF-and IL-1 [19,20]. The results of the present study revealed that the anti-inflammatory properties of TCHi in an OGD/Rdamaged NVU model were due to the reducing levels of ICAM-1 and VCAM-1 as well as cytokines. Apart from inflammation, apoptosis is another focus in the pathogenesis of brain injuries [17]. The regulation of the apoptosis process is essential to maintain the balance between cell survival and cell death, which is also important in ischemic injuries [21]. Proapoptotic proteins, such as Bax, p53, and caspase-3, were reported to take part in I/R injuries earlier [21]. The current study demonstrated that TCHi could be an effective therapy against apoptosis. It has an impact on the regulation of the apoptosis process via decreasing Bax, p53, and caspase-3 levels in an OGD/R-damaged NVU model.
Not only is BBB the key component of NVU, but also it is the most important structure to maintain cerebral homeostasis and correct neuronal function [22]. BBB integrity in the present study was reflected in TEER, tight junctions, and other ultrastructures of NVUs. TCHi was found to alleviate the BBB breakdown in the OGD/R model. Previous Evidence-Based Complementary and Alternative Medicine 9 researchers have confirmed the anti-inflammatory and antiapoptotic roles of troxerutin in chronic diseases and diabetic models [23,24]. The findings of this study revealed that TCHi, a new compound of troxerutin, had a good therapeutic effect in acute attacks such as cerebral I/R injuries.
This study still has limitations. Inflammatory cytokines and cell adhesion molecules were found to be upregulated in NVUs for the in vitro OGD/R model; however, the infiltration of peripheral leukocytes and other factors in the central nervous system could not be demonstrated. As a result, further in vivo studies should be performed to show whether TCHi had the same effects as in the in vitro model.

Conclusion
TCHi protected the main types of cells of NVUs in vivo and in vitro depending on anti-inflammation, antiapoptosis, and BBB. The data implies that TCHi is a candidate medicine to treat cerebral ischemic stroke.

Disclosure
Hóngyi Zhào and Yu Liu are co-first authors in this article.

Conflicts of Interest
The authors declare that they have no conflicts of interest.