The Protective Effect Of Annexin A2 On Trauma-Induced Blood Brain Barrier Injury And Its Mechanisms

BackgroundIn the central nervous system (CNS), blood-brain barrier (BBB) plays a critical role in creating a unique and stable environment for neuronal and synaptic functions. The blood brain barrier is composed of a tightly sealed monolayer of brain endothelial cells (BECs) that restrict entry of blood-borne molecules into the brain so that protect the brain microenvironment. Traumatic brain injury (TBI) remains a highly intractable problem with extremely limited therapeutic options. Translating molecular pathological targets into clinically effective therapies has been challenging. In recent years, vascular mechanisms have been recognized to be a central part of the complex pathophysiology of TBI. In particular, disruption in BBB integrity is widely considered to be the primary vascular event after TBI. Hence, finding novel vascular targets that can protect the BBB should be an important therapeutic approach. This is the thrust of the present translational study.Annexin A2, a Ca2+-regulated membrane-binding protein, has been shown to be integral to membrane trafficking events such as endocytosis and exocytosis. Independent experiments using calcium chelation to disrupt cell-cell junctions showed that Annexin A2 is required for the reassembly of both adherens junctions by recruiting E-cadherin to nascent adherens junctions as well as for the reassembly of tight junctions. Further more, as a member of Annexin family, Annexin A1 has been identified to be an essential endogenous regulator of BBB integrity and recombinant Annexin Al restores cell polarity, cytoskeleton integrity, and paracellular permeability through inhibition of the small G protein RhoA.All these allow us to hypothesis that Annexin A2 may act as a candidate therapeutic agent for TBI-induced BBB disruption. Our objective in this study, therefore, was to identify a new regulator of BBB function that may lead to the development of more effective treatment strategies.MethodsEmbryonic BBB permeability assay. Pregnant mice were deeply anaesthetized. Embryos were injected while still attached to the mother's blood circulation via the umbilical cord. Dye was injected using an Hmailton syringe into the embryonic liver. 10-kDa dextran was used to detect in all CNS capillaries,3 min of circulation was allowed.Postnatal and adult BBB permeability assay. Mice were deeply anaesthetized and 10-15 ml EZ-link NHS-sulfo-biotin was injected into the left heart ventricle, followed by 1% paraformaldehyde (PFA) in PBS.Controlled Cortical Impact Model for Traumatic Brain Injury (CCI). After induction of isoflurane anesthesia, a midline longitudinal incision was performed, the skin was retracted and the skull exposed. A 5.0 mm-diameter craniotomy was made in the left parietal bone midway between bregma and lambda with the medial edge 1 mm lateral to the midline. Mice were impacted at 5.0 m/s with a 40ms dwell time and 0.6mm depression using a 3mm diameter convex tip, mimicking a moderate TBI (traumatic brain injury) in humans. The bone flap was discarded, and the scalp was sutured closed. After surgery the wild-type animals were treated with saline or recombinant A2 intravenously in a double-blind manner.Transwell permeability. HBMEC were seeded on 1.0 ?m transwell inserts coated with collagen. Cells were grown to confluency and treated with different conditions. Horseradish peroxidase was added to the top chamber with a concentration of 100 ug ml-1. After 1 hr incubation, remove the lower chamber.0.5 mM of guaiacl and 0.6 mM H2O2 were added into the media obtained from lower chamber. Spectrophotometric analysis of absorbance at 490 nm provided a quantitative evaluation of the amount of horseradish peroxidase that crossed the membrane.Results1. Annexin A2 is involved in functional BBB formation in the cortex development.In brain, we found A2 expression in CNS vascular endothelial cells that have highly restrictive barrier properties. In wild type E13.5 cortex, the tracer was already confined to vessels as mature BBB, for A2-/- mice, signal still can be detected outside vessels. In E15.5 cortex,10 kDa tracer was confined to vessels both in wild type and A2-/- mice However, when we used much smaller molecular weight tracer (550 Da), the immunostaining results showed that the tracer was detected outside vessels in A2-/- mice. The same results were found in the pup (4 day) and adult mice (12 weeks).Based on the importance of junction complex in BBB, we compared their expression of wild type and A2-/-adult mice (12 weeks) in isolated cerebral cortex microvacular. For the RNA expression, there was significant decrease of tight junctions: ZO-1 and claudin-5, adherens junctions:VE-cadherin in the A2-/- mice compared to wild type. For the protein expression, there was a similar trend.2. Role of Annexin A2 on traumatic BBB dysfunctionTo test whether lacking of A2 can aggravate BBB leakage after injury, we injected small volume tracer (10 kDa) to the wild type and A2-/- mice after 24 hrs CCI surgery. They both displayed leakiness surrounding the injury area, but in A2-/- mice, there were severe barrier leakage defects. Moreover, Evans blue also showed similar results between wild type and A2-/- mice. For the junction complex, both tight junctions:ZO-1, Occludin, Claudin-5 and adherens junctions:VE-cadherin were decreased in A2-/- mice compared to wild type mice (ipsilateral side). Therefore, A2-/- mice display more BBB leakage after TBI injury.Due to Annexin A2's importance in BBB function, we next addressed whether recombinant Annexin A2 (rA2) rescue BBB leakage after TBI injury. Given different dose of rA2 (0.75 mg/kg-1.5 mg/kg) 2 hrs after CCI surgery, Evans blue extravasation was significant less than BSA control group. Specifically, even given rA2 (1.0 mg/kg) 6 hrs after CCI surgery, it still could rescue BBB leakage, suggesting that rA2 was comprehensive to the therapy time window. This conclusion was also confirmed by dextran tracer immunostaining. Because of A2's involvement in junction complex, we proposed that rA2 protective effect on BBB was related to junction complex. Indeed, from western blot examination, tight junctions:ZO-1 and adherens junctions: VE-cadherin expression were found significantly increased. Together, these findings suggest that the rA2 can rescue BBB leakage by increasing ZO-1 and VE-cadherin expression.By using a well-established method to examine the paracellular permeability of a monolayer of cells, we found that A2 knock down cell significantly increased transendothelial permeability than control group. Then we treated cultured monolayers of HBMEC with IL-1B, and detected an increase in endothelial permeability. rA2 could protect the permeability from 0.1 ?M. TBI has also been associated with cerebral hypoxic-ischemic injury, so we treated HBMEC with IL-1B plus hypoxia. Under this treatment, endothelial cell permeability was increased, in contrast, rA2 treatment significantly rescued this phenomenon. With IL-1B plus hypoxia treatment, the junctional complex expression was all significantly decreased while rA2 treatment could protect against this injury through rescuing ZO-1, VE-cadherin and Claudin-5 expression on endothelial cell membrane. The junctional complex ZO-1 and VE-cadherin cell surface localization was also demonstrated by immunostaining. Collectively, these data link A2 as a critical regulator of ZO-1 and VE-cadherin expression and localization and the protective effect of IL-1B plus hypoxia-induced vascular permeability.3. The mechanisms of Annexin A2 on traumatic BBB dysfunctionTo test whether A2 can directly bind with junctional complex, we isolated membrane fraction of HBMEC monolayers and performed immunoprecipitations using an antibody directed to A2. Eluates were probed for junctional complex. In normal condition, a mount of VE-cadherin immunoprecipitated with A2. We also found A2 directly bond with F-actin. Although A2 expression of total cell lysis did not have much change under IL-1B plus hypoxia treatment, the membrane localization of A2 was significant decreased while increased in cell nuclear. After rA2 treatment, an accumulation of A2 was found on endothelial cell membrane, which made more co-localization of A2 with VE-cadherin and F-actin. Therefore, it is plausible that the decreased expression of junctional complex on HBMEC membrane is, at least in part, mediated by loss binding with A2, and that rA2 treatment rescue this interaction by controlling the localization of A2, which in turn controls endothelial cells integrity.Annexin A2 adaptive effect on interaction between the actin cytoskeleton and cell membrane components is identified to be Rho-dependent. However, in our system, we did not see much inhibition of rA2 on RhoA activation after IL-1B plus hypoxia treatment. After IL-1B stimulation, ARF6, a known regulator of adherens protein localization, was activated while rA2 treatment inhibited this activation. Consistent with this, treatment of endothelial cells with IL-1B plus hypoxia significantly activated ARF6, while rA2 treatment almost abolished this activation. Interestingly, when transfected HBMEC with A2 siRNA, an obvious activation of ARF6 was also detected. To illustrate the relationship of ARF6 activation and junctional complex, a specific agonist of ARF6-QS11 was used. mRNA and protein levels of junctional complex were decreased as concentration of QS11 increased. Furthermore, rA2 was able to inhibit QS11-induced ZO-1 and VE-cadherin downregulation. To determine whether ARF6 cascade might be the main mechanism of rA2 protective effect on junctional complex expression, we used SecinHS, a specific inhibitor of ARF6-GTP, plus rA2 treatment compared to rA2 treatment alone in our IL-1B plus hypoxia model. We found that the rA2 treatment alone did not have better effect than the combined treatment. Together, these findings suggest that the protective effect of rA2 on junctional compelx expression after injury is mainly dependent on ARF6 signal.ConclusionWe demonstrate that A2 is required in the formation and maturation of BBB in the CNS and lacking of A2 will induce more severe BBB leakge after brain injury. Because of A2 involvement in junctional complex assembly, we propose that A2 serves as a new therapy target for brain injury-induced BBB dysregulation. Indeed, from in vivo experiment, rA2 was found to be protective for the TBI-induced BBB leakage. Also, our in vitro model suggests the potential for A2's inhibition of vascular leak without modulation of RhoA-dependent response. Interestingly, our finding also highlights the importance of the interactions between A2 and VE-cadherin, A2 and F-actin in BBB function. In addition, a novel molecular-ARF6 serves as a signaling module involved in the mechanism of rA2 protetive effect on vascular endothelial cell permeability. This approach may be particularly useful in traumatic brain injury, as the current medical therapy for CNS disease-induced BBB dysregularion is limited.