Recombinant Slit2 Reduces Surgical Brain Injury Induced Blood Brain Barrier Disruption via Robo4 Dependent Rac1 Activation in a Rodent Model

Brain tissue surrounding surgical resection site can be injured inadvertently due to procedures such as incision, retractor stretch, and electrocauterization when performing neurosurgical procedures, which is termed as surgical brain injury (SBI). Blood brain barrier (BBB) disruption due to SBI can exacerbate brain edema in the post-operative period. Previous studies showed that Slit2 exhibited vascular anti-permeability effects outside the brain. However, BBB protective effects of Slit2 following SBI has not been evaluated. The objective of this study was to evaluate whether recombinant Slit2 via its receptor roundabout4 (Robo4) and the adaptor protein, Paxillin were involved in reducing BBB permeability in SBI rat model. Our results showed that endogenous Slit2 increased in the surrounding peri-resection brain tissue post-SBI, Robo4 remained unchanged and Paxillin showed a decreasing trend. Recombinant Slit2 administered 1 h before injury increased BBB junction proteins, reduced BBB permeability, and decreased neurodeficits 24 h post-SBI. Furthermore, recombinant Slit2 administration increased Rac1 activity which was reversed by Robo4 and Paxillin siRNA. Our findings suggest that recombinant Slit2 reduced SBI-induced BBB permeability, possibly by stabilizing BBB tight junction via Robo4 mediated Rac1 activation. Slit2 may be beneficial for BBB protection during elective neurosurgeries.

to regulate endothelial function and vascular permeability. In a model of glioma cocultured endothelial cells, exogenous Slit2 pretreatment reduced blood tumor barrier (BTB) permeability which was inhibited with knockdown of Robo4 receptor 21 . Likewise, recombinant Slit2 reduced vascular hyperpermeability in a Robo4 receptor dependent manner in mouse models of retinopathy 22 and lung inflammation 23 . Furthermore, the endothelial stabilizing effect of Robo4 has been shown to be mediated by the downstream intracellular adaptor protein, Paxillin 24 .
The role of Slit2 in regulating BBB permeability after SBI has not been explored. We proposed that recombinant Slit2 administration will activate Robo4-Paxillin signal transduction pathway which will attenuate SBI induced BBB disruption and improve outcomes in a rat model.

Methods
Animals. Animal procedures followed NIH Guide for the Care and Use of Laboratory Animals. The study protocol was approved by Loma Linda University Institutional Animal Care and Use Committee (IACUC). Adult Sprague Dawley rats (n = 70), male, 280-350 g, were randomly divided into Sham or SBI groups. Animals were kept in a facility with controlled humidity and temperature, 12 h light/dark phase, and given food ad libitum.
Experimental Design and Animal Groups. Experiment 1. The expression of endogenous Slit2, Robo4 and Paxillin at various time-points following SBI was evaluated. The groups included Sham and SBI groups at 6 h, 12 h, 24 h and 72 h after injury. The peri-resection right frontal lobe samples were obtained for western blot to evaluate temporal expression of the proteins and immunohistochemistry was performed for cellular localization of Robo4 and Paxillin. Experiment 2. The BBB protective effect of recombinant Slit2 was assessed. The groups included Sham, SBI + Vehicle, SBI + Slit2 (10 µg/Kg). Rats were injected recombinant Slit2 (R and D Systems, Minneapolis, MN) or normal saline as vehicle via intraperitoneal route 1 h prior to inducing SBI 18,25 . Rats were subjected to neurological testing 24 h post-SBI and then euthanized to collect peri-resection right frontal lobe samples for western blot to evaluate BBB junction proteins and for Evans blue dye extravasation assay to evaluate BBB permeability. Experiment 3. The role of Robo4 and Paxillin in BBB protective effect of recombinant Slit2 was assessed. The groups included Sham, SBI + Vehicle, SBI + Slit2 (10 µg/Kg), SBI + Slit2 (10 µg/Kg) + Robo4 siRNA, SBI + Slit2 (10 µg/Kg) + Paxillin siRNA, SBI + Slit2 (10 µg/Kg) + Scramble siRNA. Rats were injected recombinant Slit2 (R and D Systems, Minneapolis, MN) or normal saline as vehicle via intraperitoneal route 1 h prior to inducing SBI 18,25 . The siRNA for Robo4, Paxillin and Scramble siRNA (all from Life Technologies, Grand Island, NY) was injected via intracerebroventricular (ICV) route 24 h prior to SBI. Neurological testing was performed 24 h post-SBI and peri-resection right frontal lobe samples were collected for western blot and to measure Rac1 activity. Surgical Brain Injury Rat Model. The rat surgical brain injury model was made as described previously 1,10 .
Anesthesia was induced with 4% isoflurane and kept at 2.5% during surgery. A midline incision was made on the scalp and skull was exposed. A craniotomy 5 × 5 mm was performed in right side frontal bone with margins 2 mm lateral to sagittal suture and 1 mm proximal to coronal suture. Bone flap was removed to expose the underlying dura and right frontal lobe. Right frontal lobe was partially resected following margins of the bone window. The depth of resection was extended to base of skull. After resection was completed, the cavity was irrigated with normal saline followed by intra-operative packing to ensure hemostasis. Skin incision was sutured once complete hemostasis was achieved. Sham surgery included right frontal craniotomy but dura and frontal lobe was kept intact. Post-operative analgesia was provided using buprenorphine (0.03 mg/Kg) injected subcutaneously at the end of surgery. Rats were monitored at regular intervals post-operatively and returned back to home cages after complete anesthetic recovery. Intracerebroventricular Injection. Intracerebroventricular (ICV) injection was performed under isoflurane anesthesia as previously described 26 . Rats were positioned prone on a stereotactic frame. The following coordinates were used to make a burr hole on the right parietal bone: 1 mm lateral and 1.5 mm posterior in relation to bregma, and 3.2 mm depth from the surface as described previously 26,27 . A total volume of 2 μL either Robo4 siRNA (500 pmol) or Paxillin siRNA (500 pmol) (Life Technologies, Grand Island, NY) was infused 0.5 μL/min into lateral ventricle using a 10 µL Hamilton syringe (Hamilton Co, Reno, NV) through the burr hole as described previously 26,27 . Scramble siRNA (Life Technologies, Grand Island, NY) 2 μL was infused in the negative control group. The needle was kept in position for 10 mins after infusion ended and removed over 5 mins to avoid backflow. Bone wax was used to seal the burr hole and the skin was sutured. Rats were monitored post-operatively and returned back to home cages after anesthetic recovery. The siRNAs for both Robo4 and Paxillin were reconstituted in nuclease free water for delivery as provided by the vendor and carrier was not used. To improve knockdown efficacy, two different sequences targeting Robo4 were pooled: Robo4 (i) sense, 5′-GACUACGAAUUCAAAGUGAtt-3′; antisense, 5′-UCACUUUGAAUUCGUAGUCtt-3′; (ii) sense, 5′-CAGCCUCAGUAGUCGACUAtt-3′; antisense, 5′-UAGUCGACUACUGAGGCUGct-3′. Additionally, two different sequences targeting Paxillin were pooled: Paxillin (i) sense, 5′-GCUUCAUUGUCAGAUUUCAtt-3′; antisense, 5′-UGAAAUCUGACAAUGAAGCca-3′; (ii) sense, 5′-GGACAACCCUACUGUGAAAtt-3′; antisense, 5′-UUUCACAGUAGGGUUGUCCgt-3′.
Neurological Evaluation. The modified Garcia test was performed to evaluate sensorimotor deficits 24 h after SBI in a blinded manner as previously described 10,28,29 . Briefly, the test evaluated seven parameters: activity in cage, proprioception, whisker touch response, symmetry of limbs, turning, outstretching of forepaws, and climbing. Each parameter received a score ranging from 0 or 1 upto 3 to get a total score ranging from 3 to 21. Higher scores indicated better performance.
Evans Blue Dye Extravasation. Evans blue dye extravasation assay was performed 24 h after surgery to assess BBB permeability as previously described 1,26 . Briefly, Evans blue dye (2%; 5 mL/kg) was given by intraperitoneal injection and allowed to circulate for 4 h after injection. During sacrifice, transcardial perfusion of phosphate buffered saline (PBS) was performed after which the brain samples were removed, snap frozen in liquid nitrogen and stored at −80 °C until use. The peri-resection right frontal samples was homogenized in PBS (1 mL/300 mg) and then centrifuged at 14,0000 rpm for 30 mins after which the supernatant was collected. An equal amount of trichloroacetic acid (50%) was added to 500 μL of the supernatant and allowed to incubate overnight at 4 °C. The supernatant was centrifuged using the same parameters next day. The extravasation of Evans blue dye was measured at 620 nm with a spectrophotometer. The extravasated Evans blue in brain samples was quantified using a standard curve and expressed as micrograms per gram of brain tissue as previously described 30-32 . Western Blot Analysis. Western blot protocol was followed as described previously 33,34 . During sacrifice, rats were perfused through the heart with PBS and brain samples were obtained. Ripa lysis buffer (Santa Cruz Biotechnology, Dallas, TX) was used to homogenize the right frontal lobe samples and then centrifuged at 14,000 g for 30 mins at 4 °C. Supernatants were collected and protein concentration was measured. Samples (50 µg) were loaded to sodium dodecyl sulfate polyacrylamide gel for electrophoresis then transferred to nitrocellulose membrane. Following primary antibodies were applied on to membranes and incubated at 4 °C overnight: anti-Slit2 (1:500), anti-Robo4 (1:3,000), anti-Paxillin (1:200) (all from Santa Cruz Biotechnology, Dallas, TX), anti-occludin (1: 50,000) and anti-claudin3 (1:1,000) (both Abcam, Cambridge, MA), and anti-VE cadherin (1:200) (Santa Cruz Biotechnology, Dallas, TX). Membranes were incubated with anti-β Actin (1:2,000) (Santa Cruz Biotechnology, Dallas, TX) as loading controls. Secondary antibodies (1:2,000) (Santa Cruz Biotechnology, Dallas, TX) were applied on to membranes for 1 h at room temperature. The ECL Plus Chemiluminescence reagent (Amersham Biosciences, Arlington Heights, IL) was applied to visualize the bands. Density of the bands was quantified using Image J software (National Institutes of Health, Bethesda, MD). Relative density of bands was calculated as ratio to β-actin and then normalized to average of Sham group as described previously 30 . Detailed information of the antibodies used for western blot experiments are listed in Supplementary Table 1. Immunohistochemistry. Briefly, rats were perfused with PBS and formalin after which the brain samples were collected and stored in formalin. Using a cryostat (CM3050S; Leica Microsystems, Bannockburn, IL), 10 µm thick sections were obtained. Immunofluorescence staining procedure was performed as described previously 33,34 . The following primary antibodies were applied onto sections and incubated at 4 °C overnight: anti-Robo4 (1:100) or anti-Paxillin (1:100) (both from Santa Cruz Biotechnology, Dallas, TX) and co-localized with anti-von willibrand factor (vWF) (1:100) (Santa Cruz Biotechnology, Dallas, TX), anti-glial fibrillary acidic protein (GFAP) (1:100) or anti-neuronal nuclear protein (NeuN) (1:100) (both from Abcam, Cambridge, MA). Next, appropriate FITC-and Rhodamine Red-conjugated secondary antibodies (1:100) (Jackson Immuno Research, West Grove, PA) were applied at room temperature for 2 h. A fluorescence microscope (Olympus BX51) was used to visualize the sections. Rac1 Activity Assay. A Rac1 activity kit (Cell Biolabs, San Diego, CA) was used to evaluate Rac1 activity as described previously 35,36 . PAK1-PBD agarose beads was added to right frontal lobe samples and incubated for 1 h at 4 °C to pull-down active Rac1 in samples. Samples were centrifuged to isolate the beads. Sample buffer was added to the bead samples followed by electrophoresis using 10% polyacrylamide gel and nitrocellulose membrane transfer. Anti-Rac1specific antibody from the kit was applied to membranes and incubated at 4 °C overnight. Films were developed after applying appropriate secondary antibody for 2 h at room temperature to detect Rac1-GTP.
Statistical Analysis. Data were expressed as mean ± SEM, and analyzed using Sigma Plot 10.0 and Sigma Stat 3.5 (Systat Software, San Jose, CA). One-way ANOVA for multiple comparisons and Student-Newman-Kuels post-hoc test was used to analyze differences among groups. P value less than 0.05 was taken as statistically significant.

Results
Time course expression of Slit2 post-SBI. Slit2 expression in the peri-resection right frontal lobe samples was measured using western blot analysis of brain samples collected at 6 h, 12 h, 24 h, 72 h after SBI. There was an increase in the expression of Slit2 at all time points post-SBI compared to sham group (p < 0.05; Fig. 1A).

Recombinant Slit2 administration preserved BBB tight junction protein expression 24 h post-SBI.
The BBB junction proteins occludin, claudin3 and VE cadherin in the peri-resection right frontal samples was evaluated using western blot analysis 24 h post-SBI. Occludin and claudin3 expression was reduced significantly post-SBI compared to sham group (p < 0.05; Fig. 4A and B, respectively). Recombinant Slit2 administered SBI group had significantly increased expression of occludin compared to vehicle group (p < 0.05) and a trend to increase the expression of claudin3 post-SBI. However, the expression of VE cadherin was not significantly different between sham or SBI groups at 24 h (p > 0.05; Fig. 4C).
Recombinant Slit2 administration attenuated BBB permeability and neurological deficits 24 h post-SBI. The permeability of BBB was evaluated using Evans blue dye extravasation assay at 24 h after SBI.
Rats subjected to SBI had significantly increased extravasated dye in the peri-resection right frontal tissue compared to sham group (p < 0.05; Fig. 5A). Recombinant Slit2 administration reduced significantly the extravasation of Evans blue dye post-SBI compared to vehicle group (p < 0.05). Recombinant Slit2 administered SBI group had significantly lower neurological deficits, which was observed as increased neurological scores evaluated by modified Garcia test 24 h post-SBI (p < 0.05; Fig. 5B).

Discussion
This study examined the BBB protective effects of recombinant Slit2 administration in a rat SBI model. Our findings showed that recombinant Slit2 reduced SBI induced BBB permeability which was associated with the preservation of endothelial tight junction protein expression. Robo4 receptor and Paxillin was expressed by endothelial cells at the right frontal peri-resection tissue post-SBI. SBI rats that received recombinant Slit2 showed increased Rac1 activity post-SBI that was reversed with Robo4 siRNA and Paxillin siRNA administration along with recombinant Slit2. Our findings suggest that recombinant Slit2 attenuated BBB disruption after SBI, possibly via stabilizing tight junction protein through Robo4-Paxillin dependent activation of Rac1 after SBI.
We first evaluated the endogenous Slit2 expression at various time-points post-SBI. Our findings showed that Slit2 started increasing at 6 h following SBI and remained elevated till 72 h following injury. Slit2 has been shown to be expressed by various cell types in the brain including neurons, astrocytes, and endothelial cells [37][38][39] . Previous studies from our lab and others have shown that Slit2 increased in response to an insult in the brain including traumatic brain injury (TBI) 17 , SBI 18 and cerebral ischemia 39 . Additionally, our previous study showed that endogenous Slit2 had a protective function after SBI 18 . These finding suggest that Sli2 has protective role following brain injury. We therefore sought to examine whether recombinant Slit2 administration would augment protection after SBI. Specifically, in this study we explored the BBB specific protective function of recombinant Slit2 in a rat SBI model.
Slit2 is known to function by binding to Robo receptors. Among the Robo family receptors, Robo4 is known to be the endothelial specific receptor for Slit2 ligand 14,20 . Robo4 receptor has been demonstrated to be expressed by endothelial cells 20 including human brain microvascular endothelial cells 21 . Consistent with these findings, we observed that Robo4 was expressed by endothelial cells at the perisurgical site in SBI rats. Interestingly, we also found that Robo4 was expressed by neurons but not astrocytes surrounding the perisurgical site. Additionally, our results showed that the expression of Robo4 evaluated 24 hours after SBI did not change compared to sham group. In a transient cerebral ischemia rat model, Robo4 weakly increased at day 3 and continued to increase progressively at days 14 and 28 after reperfusion and was expressed by reactive astrocytes in the CA1 region of hippocampus 39 . It is likely that expression of Robo4 is induced in the long term after injury during reactive astrogliosis. It is also possible that Robo4 may be differentially expressed in varying pathologies. Previous study showed that the expression of Robo4 was upregulated in human glioma tissues compared to normal brain tissue 21 . Furthermore, Robo4 was shown to be abundantly expressed by endothelial cells in various tumor samples 40 . In a tumor environment, Robo4 has been implicated to regulate the migration of endothelial cells during tumor angiogenesis 40 . Since Robo receptors are known to mediate their function via activation of intracellular signal transduction pathways, we suggest that in the presence of recombinant Slit2 the endothelial protective Robo4 signaling pathway gets activated despite no change in receptor expression after SBI. The regulation of Robo4 expression after SBI remains to be further explored.
Paxillin is an intracellular adaptor protein that binds to structural and signaling molecules 24,41 . It functions as a scaffold protein to which signaling molecules dock and activate intracellular signal transduction pathways 42 . Previous study showed that Paxillin was a critical downstream effector of Robo4 receptor in endothelial cells 24 . Furthermore, Paxillin was shown to play a key role in endothelial barrier enhancement in human lung microvascular endothelial cells 43 . We therefore investigated if Paxillin played a role in BBB stabilization by recombinant Slit2. First, we evaluated the expression of Paxillin after SBI. We observed that Paxillin was expressed by endothelial cells and astrocytes at the perisurgical site but not by neurons. Additionally, we observed a decreasing trend in the expression of Paxillin after SBI despite a strong expression of Paxillin noted in astroctyes. The decrease in Paxillin after SBI therefore may be more selective for the vasculature and neurons. Since we used whole brain samples taken from the perisurgical site which includes all cell types for western blot experiments, this may partly explain the decreasing trend in Paxillin expression observed after SBI despite a strong expression seen in astrocytes. Various factors such as extracellular matrix proteins, tyrosine kinases, cytokines 44 , growth factors 45 , lipopolysaccharide 41 have been reported to regulate Paxillin. These factors have been shown to increase after SBI and may negatively regulate Paxillin following injury 9,10,34,46 .
Although primarily involved in regulating migration of developing axons and neurons 16,19 , recent studies have demonstrated that Slit2 can exert its effects on other cells types including endothelial cells and regulate endothelial barrier integrity 21,47,48 . Several studies have shown that Slit2 protected endothelial integrity. Slit2 maintained endothelial barrier stability of human lung lymphatic endothelial cells exposed to HIV in vitro 47 . Additionally,   The vascular protective effect of Slit2 has been implicated to be dependent on Robo4. Slit2 was found to inhibit VEGF induced retinal hyperpermeability in Robo4+/+ mice but not in Robo4 null mice suggesting that Slit2-dependent inhibition of VEGF induced permeability was mediated via Robo4 activation 22 . Likewise, VEGF induced vascular permeability was attenuated by Slit2 in Robo4+/+ but not in Robo4−/− endothelial cells 49 . Likewise, Slit2 reduced Andes virus induced pulmonary microvascular endothelial cell permeability dependent on Robo4 in vitro 48 . In accordance with previous studies, SBI rats that received recombinant Slit2 had decreased extravasation of albumin-bound Evans blue dye into peri-resection brain tissue post-SBI suggesting that Slit2 protected against SBI-induced BBB hyperpermeability.
The BBB interendothelial junction proteins are important components of BBB that maintain integrity of the BBB 50 . Tight junction proteins such as occludin and claudin3 maintain barrier function of the BBB, and disruption of tight junction proteins increases endothelial barrier permeability after an insult to the brain 1,51 . Various factors such as inflammation, oxidative stress and proteases can degrade tight junction proteins after brain injury 51,52 . Previous studies show that tight junction proteins were degraded in the perisurgical site thereby increasing BBB permeability and consequently worsening brain edema in SBI rodent models 1,10 . The endothelial protective function of Slit2 was shown to be associated with stabilization of inter-endothelial junctions in vitro 48 . In addition, transgenic overexpression of Robo4 by glioma cocultured endothelial cells reduced BTB permeability by preserving tight junction protein expression in vitro 21 . We observed that recombinant Slit2 administration partially increased the expression of tight junction proteins after SBI suggesting that Slit2 regulates the expression ANOVA, SNK. N = 4/group. *p < 0.05 compared to Sham, † p < 0.05 compared to SBI + Vehicle, @ p < 0.05 compared to SBI + Slit2 10 μg/Kg, # p < 0.05 compared to SBI + Slit2 + Robo4 siRNA, $ p < 0.05 compared to SBI + Slit2 + Paxillin siRNA. The full length blots for western blot pictures shown in panels A and B, and Rac1 activity assay shown in panel C are presented in Supplementary Figure S4. of endothelial tight junction proteins. However, we did not observe any change in expression of adherens junction protein VE cadherin 24 hours after SBI with or without Slit2 administration.
The small GTPase protein Rac1 is an important mediator known to stabilize tight junction and adherens junction complexes that maintain endothelial barrier integrity 35,53 . Robo4 overexpression in endothelial cells was shown to activate Rac1 in vitro whereas, knockdown of Robo4 reduced Rac activation 54 . We speculated that recombinant Slit2 binds to Robo4 receptors on endothelial cells and regulates the expression of tight junction proteins possibly by affecting Rac1 activation. We therefore investigated whether Slit2 regulates Rac1 activity dependent on Robo4. We examined the role of Robo4 and its downstream mediator Paxillin in recombinant Slit2 mediated regulation of Rac1 activity. The Paxillin Interaction Motif (PIM) in the cytoplasmic tail of Robo4 directly interacts with Lim domain of Paxillin 24 , and this interaction is increased when Slit2 is present. There are at least two possible ways by which Paxillin can initiate Rac1 activation. First, Paxillin can recruit Rac-specific GEF, bPIX which increases Rac1 activity 45 . Second, binding of Slit2 ligand to Robo4 increases the interaction of Robo4 with Paxillin and Arf-GAP complex 24 . The Arf-GAP complex inactivates Arf6 which is a small GTPase protein that induces the internalization of Rac1 55,56 . In the presence of Slit2, increased interaction between Robo4-Paxillin inactivates Arf6 thereby restoring Rac1 activity. In accordance, we observed that recombinant Slit2 administration increased Rac1 activity after SBI which was reversed with siRNA knockdown of either Robo4 or Paxillin. Based on these findings, we infer that Slit2 preserved BBB integrity after SBI possibly by Robo4-Paxillin mediated activation of Rac1 which partly stabilized the tight junction proteins. Of note, we observed that Paxillin knockdown appeared to influence Robo4 expression after SBI although it was not significant. The off-target effects of siRNAs cannot be completely ruled out. However, overall the knockdown of either Robo4 or Paxillin reversed Rac1 activity increased by recombinant Slit2 suggesting that Robo4 and Paxillin were involved in Rac1 activation with recombinant Slit2.
Our study had some limitations. First, we did not provide direct evidence to show that Robo4 and Paxillin interaction increases upon recombinant Slit2 administration. Previous study showed that the interaction between Paxillin and Robo4 receptor is increased by Slit2 24 . Although we do not have direct evidence to show Robo4 and Paxillin interaction increased Rac1 activity upon Slit2 administration, our results showed that recombinant Slit2 administration increased Rac1 activity which was reversed with Robo4 and Paxillin knockdown. This suggests that Robo4 and Paxillin played a role in Rac1 activation in the presence of recombinant Slit2. Second, Rac1 may be among one of the factors that regulate BBB stability after SBI. Therefore, despite an increase in Rac1 activity with recombinant Slit2 administration, BBB junction stability and leakage was partly reversed after SBI. Third, it is possible that Robo4 may modulate BBB integrity through inhibition of matrixmetalloproteinase-9 (MMP-9) activity. Previous study showed that knockdown of Robo4 increased BTB permeability by upregulating MMP-9 in glioma endothelial cells in vitro 21 . It is possible that recombinant Slit2 modulates MMP activity dependent on Robo4 to preserve BBB junction integrity. Additionally, other downstream molecules may be involved in Robo4 signaling pathway. For instance, the phosphorylation of Src and Erk1/2 were upregulated with knockdown of Robo4 in vitro 21 . Likewise, abelson tyrosine kinase (Abl) and its substrate enabled (Ena) interact with the cytoplasmic tail of Robo4 14,49 . Although, we did not evaluate alternate pathways, it is possible that Robo4 may modulate various downstream factors including Paxillin and Rac1 that may have contributed to the endothelial protective effects of recombinant Slit2. Lastly, we did not evaluate effects of Slit2 on other cell types apart from the BBB protective function. Recombinant Slit2 possibly crosses the BBB once it is disrupted following SBI and can therefore exert effects on other cell types in the brain including neurons that have been reported to express Robo receptors. In this study our objective was to evaluate the BBB protective effects of Slit2, therefore Slit2 was administered by intraperitoneal systemic route and the extravascular effects were not evaluated. We hypothesized that recombinant Slit2 interacts with endothelial Robo4 receptors thereby protecting against BBB permeability. Further studies are required to investigate the extravascular function of Slit2.
Overall, this study showed that recombinant Slit2 reduced BBB permeability by stabilizing endothelial tight junction proteins after experimental SBI possibly via Robo4-Paxillin dependent Rac1 activation.