Search

Few data are available on the involvement of brain microvascular endothelial cells (BMECs) in excitotoxic neonatal brain lesions. Therefore, we developed an original approach for investigating mouse-derived BMECs in vitro. We hypothesized that newborn and adult BMEC cultures would show age-related differences in phenotype and sensitivity to glutamate. Expression of the monocarboxylate transporter, MCT1, was higher in neonatal than in adult BMECs, whereas expression of the glucose transporter, GLUT1, was higher in adult than in neonatal BMECs that overexpressed the N-methyl-D-aspartate receptor NR1 subunit (NMDAR1) compared with adult BMECs. The ability of neonatal and adult BMECs to be activated by glutamate was confirmed through intracellular calcium ([Ca2+]i) recording. The glutamate-induced [Ca2+]i increase was blocked by the selective NMDAR antagonist, MK-801. Significant glutamate-evoked concentration-dependent release of tissue-type plasminogen activator (t-PA) and matrix metalloproteinases (MMPs) activities was found in supernatants of neonatal, but not in adult BMECs. The glutamate-mediated release of t-PA, MMP-2, and MMP-9 proteolytic activities in neonatal BMECs was blocked by MK-801. Conceivably, this protease release from neonatal BMECs may participate in neonatal brain lesions.

Stroke results from the occlusion of a cerebral artery by a fibrin clot. Currently, the only approved treatment for stroke is the use of tPA (tissue-type Plasminogen Activator) as a thrombolytic agent 1. Although tPA is beneficial in the vascular compartment, tPA is deleterious in the cerebral parenchyma by modifying properties of the NMDA receptor. Indeed, tPA cleaves the NR1 subunit of the NMDA receptor leading to a potentiation of the NMDA-induced calcium influx 2. Based on these observations, we have investigated whether neuroserpin (NS), a brain-derived tPA inhibitor 3, could prevent this deleterious pro-excitotoxic activity of tPA. By using cultured cortical neurons, we have tested whether recombinant NS could influence excitotoxicity induced by NMDA or AMPA, two glutamatergic agonists. Although NS failed to alter AMPA-induced necrosis, it displayed a dose-dependent neuroprotective activity (up to 50%) against NMDA receptor-mediated neuronal death. By using calcium videomicroscopy imaging, we evidenced that NS decreases NMDA-induced calcium influx in neurons. Moreover, preliminary results show that NS reduces the lesion induced by the striatal injection of NMDA. These data suggest that promoting NS expression in the brain of stroke patients could prevent the deleterious effect of tPA in the brain parenchyma (potentiation of excitotoxic neuronal death) without affecting the beneficial thrombolytic activity of tPA in the vascular compartment. To address this question, we are investigating the regulation of NS expression by using luciferase gene reporter, real time PCR and immunoblotting to identify new ways to control the production of endogenous NS.

Despite uncontroversial benefit from its thrombolytic activity, the documented neurotoxic effect of tissue plasminogen activator (tPA) raises an important issue: the current emergency stroke treatment might not be optimum, if exogenous tPA can enter the brain and thus add to the deleterious effects of endogenous tPA within the cerebral parenchyma. Here, we aimed at determining whether vascular tPA crosses the blood-brain barrier (BBB) during cerebral ischemia, and if so, by which mechanism. We have first evidenced that brain lesions induced by intra-striatal injection of NMDA were enhanced either when tPA was co-infused in the striatum or when intravenously injected. In addition, co-infusion of PAI-1 with NMDA in the striatum prevented the deleterious effect of intravenous injection of tPA. No BBB alteration was observed in any of these conditions, as assessed by confocal microscopy analysis of FITC-Dextran (77 kDa) extravasation. Altogether, these data suggest that tPA (69 kDa) has to cross the intact BBB to potentiate neuronal death within the parenchyma. By using biotinylated tPA, we have shown by immunohistochemistry, that tPA crossed the BBB in vivo in control and NMDA-injected animals. Moreover, tPA can be detected in the cerebrospinal fluid after its intravenous injection. In order to characterize the mechanism of this passage, we used a cellular model of BBB. Although tPA did not influence the integrity of the BBB by itself, tPA was capable to cross the BBB. Its passage was blocked at 4 C and was saturable, suggesting a transendothelial and receptor–mediated mechanism. As the low-density lipoprotein receptor related protein (LRP) and mannose receptor have been already shown to bind tPA, we investigated whether one of these two receptors could be involved in this passage. Although mannose had no effect on the passage of tPA, RAP (receptor–associated protein, an antagonist of LRP) blocked the passage of tPA. We next investigated whether oxygen and glucose deprivation (OGD) could influence the mechanism of the passage of tPA. OGD led to an exacerbation of tPA passage, and switched the mechanism to a LRP-independent process. These data suggest two mechanisms by which tPA can cross the BBB, a moderate receptor-mediated transcytosis in control conditions and an exacerbated and unsaturable passage involving an unspecific transcellular pathway after ischemic conditions. Preventing the interaction of tPA with LRP could thus be an interesting strategy to block the deleterious effect of tPA in vivo, but only as long as the integrity of the BBB is not altered. These data show the importance of taking the side effects of blood derived tPA into account, and offer bases to improve the current thrombolytic strategy for stroke treatment.

Over the last decade, tissue type plasminogen activator (tPA) has been implicated in a variety of brain functions. It is secreted from growth cones or extending neurites, modulates neurite outgrowth and promotes neuronal migration in vitro. tPA has also been involved in glutamatergic-dependent processes such as synaptic plasticity and long term potentiation. For example, tPA deficient mice display alterations in learning paradigms and memory processes. Similarly, tPA has been shown to play an important role in the pathogenesis of seizures, multiple sclerosis, trauma and ischemic brain injury 1. Thrombolysis by tissue type plasminogen activator (tt-PA) is currently the only approved treatment in acute ischemic stroke (NINDS, 1995)2. However, experimental models suggest that rt-PA could have deleterious effects by potentiating ischemic neuronal death. We have previously demonstrated that rt-PA promotes excitotoxic neuronal death by interacting with and cleaving the N-terminus of the NMDA receptor NR1 subunit, at arginine 260, thereby augmenting NMDA receptor-signalling 3-4. DSPA is a related plasminogen activator derived from bat saliva and lacks the Kringle2 domain. More recently, DSPA administered 3 to 9 hours after acute ischemic stroke in man, has been shown to produce high reperfusion rates and an improved clinical outcome with a favorably low bleeding risk 5. In this study, we demonstrate that DSPA, unlike rt-PA, neither interacts with nor cleaves the NR1 subunit. These data are consistent with the observation that, in contrast to rt-PA, DSPA does not potentiate NMDA-induced neuronal death in murine models 6. They provide a molecular basis that possibly explains why DSPA has a more favorable profile compared to rt-PA in the treatment of acute ischemic stroke.

Tissue-type plasminogen activator (tPA) is thought to be involved in several brain functions and dysfunctions. Its vascular thrombolytic activity led to its use for ischemic stroke treatment. However, increasing evidence suggests that tPA potentiates excitotoxic neuronal death. Accordingly, we have recently shown that tPA can directly modulate N-methyl-D-aspartate (NMDA) receptor signalling by cleaving the arginine 260 within the amino-terminal domain of the NR1 subunit (ATD-NR1). The deleterious effects of tPA during excitotoxic neuronal death suggests that blocking its effect on NMDA-signalling might improve stroke treatment without hindering its beneficial fibrinolytic action. To validate this postulate, and to investigate the relevance of tPA-mediated control of NMDA receptor signalling in vivo, we developped a protocol of immunization in mice against ATD-NR1 including the interaction domain with tPA. Complete Freund's adjuvant (CFA) or recombinant ATD-NR1 (amino acids 19-371) dissolved in CFA were injected intraperitonally once a week during four weeks. By immunoblotting, we show that sera harvested from immunized mice recognized recombinant ATD-NR1, while sera from control mice (CFA only) did not. We then postulated that the antibodies produced in the immune response might have a therapeutic value by physically preventing the interaction between tPA and NR1. We have thus performed excitotoxic lesions by injecting NMDA into the striatum of control and immunized mice, with or without intravenous injection of tPA (1 mg/kg). In CFA-injected animals, while NMDA alone led to an excitotoxic lesion of around 13 mm3, tPA, when injected intravenously, potentiated twice the lesions (29.5 mm3). In contrast, intravenous tPA was unable to potentiate NMDA-induced excitotoxic injury in ATD-NR1-vaccinated mice. No difference was observed between NMDA-induced lesion in naive, CFA-, or ATD-NR1- vaccinated mice. We then addressed the relevance of this tPA/NMDA receptor interaction in brain physiology. Since tPA is thought to modulate some cognitive functions, we tested the immunized mice in a hippocampal-dependent spatial memory task (Y maze) that does not require previous learning. Locomotor activity was not different between naive, CFA-injected and ATD-NR1-immunized mice. However, spatial memory was clearly impaired in ATD-NR1-immunized mice as compared to control or CFA-injected mice. This deficit was transient, since it disappeared 5 weeks after the last injection. Mice were also tested in an amygdala-dependent social discrimination task. Social memory was also impaired in immunized mice as compared to control or CFA-injected mice. The binding of endogenous tPA to NR1 is thus physiologically relevant, suggesting that the proteolytic activity of tPA is critical for some normal functions of NMDA receptors in the adult brain. In conclusion, this is the first in vivo demonstration that the ability of tPA to potentiate NMDA receptor signalling is critical, both during physiological and pathological brain conditions. Moreover, the immunization protocol developped here opens therapeutic and fundamental scientific interesting perspectives.

Given the potent anti-inflammatory effects exerted by glucocorticoids under various pathological situations, their use has been proposed to protect the brain against cerebral ischemia, neuro-inflammatory diseases and trauma. However, the efficiency of these treatments remains a matter of debate, since both neuroprotective and neurotoxic effects have been observed depending on the experimental models used (Abraham et al., J Neuroendocrinol., 13: 749, 2001). First we observed that purified neuron and astrocyte populations both express the mRNA encoding for the glucocorticoid receptor (as determined by RT PCR), suggesting that both cell type might respond to glucocorticoid. In order to further understand the role of glucocorticoids in brain tissues, we have tested the effect of dexamethasone (a potent glucocorticoid) in a model of mixed murine cultures of neurons and glia subjected to N-methyl-D-aspartate (NMDA)-induced excitotoxicity. We observed that pre-treatment of the cultures with dexamethasone significantly reduced NMDA-induced neuronal necrosis. We also described that dexamethasone (dose and time dependently) induced within 12–24 hours of treatment a significant increase in both the synthesis and secretion of type 1 plasminogen activator inhibitor (PAI-1) in astrocyte cultures. Previous experiments in the laboratory have shown in the same paradigm of NMDA-induced toxicity that tPA markedly increases the deleterious action of NMDA and that this potentiating action of tPA is abolished by PAI-1 (Buisson et al., FASEB J, 12: 1683, 1998). Since dexamethasone led to a significant decrease in t-PA activity (as assessed by zymography assay), our results suggest that glucocorticoids might exert a neuroprotective activity against NMDAR-mediated toxicity by interacting with the t-PA/PAI-1 axis.