TAT peptide at treatment-level concentrations crossed brain endothelial cell monolayer independent of receptor-mediated endocytosis or peptide-inflicted barrier disruption.

The peptide domain extending from residues 49 to 57 of the HIV-1 Tat protein (TAT) has been widely shown to facilitate cell entry of and blood-brain barrier (BBB) permeability to covalently bound macromolecules; therefore, TAT-linked therapeutic peptides trafficked through peripheral routes have been used to treat brain diseases in preclinical and clinical studies. Although the mechanisms underlying cell entry by similar peptides have been established to be temperature-dependent and cell-type specific and to involve receptor-mediated endocytosis, how these peptides cross the BBB remains unclear. Here, using an in vitro model, we studied the permeability of TAT, which was covalently bound to the fluorescent probe fluorescein isothiocyanate (FITC), and evaluated whether it crossed the "in vitro BBB", a monolayer of brain endothelial cells, and whether the mechanisms were similar to those involved in TAT entry into cells. Our results show that although TAT crossed the monolayer of brain endothelial cells in a temperature-dependent manner, in contrast to the reported mechanism of cell entry, it did not require receptor-mediated endocytosis. Furthermore, we revisited the hypothesis that TAT facilitates brain delivery of covalently bound macromolecules by causing BBB disruption. Our results demonstrated that the dose of TAT commonly used in preclinical and clinical studies did not exert an effect on BBB permeability in vitro or in vivo; however, an extremely high TAT concentration caused BBB disruption in vitro. In conclusion, the BBB permeability to TAT is temperature-dependent, but at treatment-level concentrations, it does not involve receptor-mediated endocytosis or BBB disruption.

Blood-Brain Barrier Disruption in Preclinical Mouse Models of Stroke Can Be an Experimental Artifact Caused by Craniectomy.

The pathophysiological features of ischemia-related blood-brain barrier (BBB) disruption are widely studied using preclinical stroke models. However, in many of these models, craniectomy is required to confirm arterial occlusion via laser Doppler flowmetry or to enable direct ligation of the cerebral artery. In the present study, mice were used to construct a distal middle cerebral artery occlusion (dMCAO) model, a preclinical stroke model that requires craniectomy to enable direct ligation of the cerebral artery, or were subjected to craniectomy alone. dMCAO but not craniectomy caused neurodegeneration and cerebral infarction, but both procedures induced an appreciable increase in BBB permeability to Evans blue dye, fluorescein, and endogenous albumin but not to 10 kDa dextran-FITC, leading to cerebral edema. Using rats, we further showed that BBB disruption induced by craniectomy with no evidence of dural tearing was comparable to that induced by craniectomy involving tearing of the dura. In conclusion, our data demonstrated that craniectomy can be a major contributor to BBB disruption and cerebral edema in preclinical stroke models. The implications of this experimental artifact for translational stroke research and preclinical data interpretation are discussed.

Drug-Induced Hyperglycemia as a Potential Contributor to Translational Failure of Uncompetitive NMDA Receptor Antagonists.

Hyperglycemia is a comorbidity in 60-80% of stroke patients; nevertheless, neuroprotective drugs like NMDA receptor (NMDAR) antagonists are typically assessed in normoglycemic animals at the preclinical stage before they are approved to enter clinical trials. Interestingly, as a possible explanation for the translational failure of NMDAR antagonists, it was recently reported that stroke occurring during nighttime causes smaller infarctions in rodents and therefore has a smaller window for neuroprotection. To investigate why stroke occurring during different circadian phases confers a difference in severity, we reanalyzed the published source data and found that some mice that were used in the daytime have higher blood glucose than mice that were used in the nighttime. We then repeated the experiments but found no difference in blood glucose concentration or infarct volume regardless of the circadian phase during which stroke occurs. On the other hand, induction of hyperglycemia by glucose injection reproducibly increased stroke severity. Moreover, although hyperglycemia increases infarction volume, which presumably would provide a larger window for neuroprotection, uncompetitive NMDAR antagonists were unexpectedly found to exacerbate stroke outcome by worsening hyperglycemia. Taken together, our new data and reanalysis of the published source data suggested that blood glucose during stroke, rather than the circadian phase during which stroke occurs, affects the size of the ischemic infarction; moreover, we have revealed drug-induced hyperglycemia as a potential reason for the translational failure of uncompetitive NMDAR antagonists. Future trials for this class of neuroprotective drugs should monitor patients' blood glucose at enrollment and exclude hyperglycemic patients.

Stroke injury induced by distal middle cerebral artery occlusion is resistant to N-methyl-d-aspartate receptor antagonism in FVB/NJ mice.

Although N-methyl-d-aspartate receptor (NMDAR) antagonism has been shown to have a neuroprotective effect in many preclinical stroke models, the efficacy of this antiexcitotoxicity strategy in clinical trials in stroke patients has been disappointing. Interestingly, it has been reported that NMDAR antagonism is not neuroprotective in C57BL/6 mice subjected to distal middle cerebral artery occlusion (dMCAO), supporting the notion that whether these treatments are neuroprotective depends on the type of cerebral ischemia. However, because C57BL/6 mice are inherently resistant to excitotoxicity, the reported lack of neuroprotection could also be explained by the difference in the mouse strain studied rather than the stroke model used. Here we examined the neuroprotective efficacy of NMDAR antagonism in FVB/NJ mice, an excitotoxicity-prone mouse strain, subjected to dMCAO. Although C57BL/6 mice are known to have an excitotoxicity-resistant genetic background and FVB/NJ mice are known to have an excitotoxicity-prone genetic background, the infarct volume and density of neurodegenerating neurons were similar in the two mouse strains following dMCAO. In addition, none of the antiexcitotoxicity agents studied, including the canonical NMDAR antagonist MK801 and the therapeutic peptides Tat-NR2B9c and L-JNKI-1, protected the FVB/NJ mouse brain against ischemic damage induced by dMCAO. In conclusion, our data demonstrated that FVB/NJ mice are no more susceptible to cerebral ischemia than C57BL/6 mice and that NMDAR antagonism is ineffective in mice, even in an excitotoxicity-prone strain, subjected to dMCAO.