Zika virus (ZIKV) has been associated to central nervous system (CNS) harm, and virus was detected in the brain and cerebrospinal fluids of microcephaly and meningoencephalitis cases. However, the mechanism by which the virus reaches the CNS is unclear. Here, we addressed the effects of ZIKV replication in human brain microvascular endothelial cells (HBMECs), as an in vitro model of blood brain barrier (BBB), and evaluated virus extravasation and BBB integrity in an in vivo mouse experimental model. HBMECs were productively infected by African and Brazilian ZIKV strains (ZIKVMR766 and ZIKVPE243), which induce increased production of type I and type III IFN, inflammatory cytokines and chemokines. Infection with ZIKVMR766 promoted earlier cellular death, in comparison to ZIKVPE243, but infection with either strain did not result in enhanced endothelial permeability. Despite the maintenance of endothelial integrity, infectious virus particles crossed the monolayer by endocytosis/exocytosis-dependent replication pathway or by transcytosis. Remarkably, both viruses' strains infected IFNAR deficient mice, with high viral load being detected in the brains, without BBB disruption, which was only detected at later time points after infection. These data suggest that ZIKV infects and activates endothelial cells, and might reach the CNS through basolateral release, transcytosis or transinfection processes. These findings further improve the current knowledge regarding ZIKV dissemination pathways.

Human infection by different flaviviruses may cause severe neurologic syndromes, through pathogenic mechanisms that are still largely unknown. Japanese encephalitis virus (JEV), West Nile virus (WNV), Zika virus (ZIKV), yellow fever virus (YFV), dengue virus (DENV), and tick-borne encephalitis virus (TBEV) are believed to reach the central nervous system by a hematogenous route, upon crossing the blood-brain barrier. Although the disruption of BBB during flavivirus infection has been largely evidenced in experimental models, the relevance of BBB breakdown for virus entering the brain was not completely elucidated. In vitro models of BBB had demonstrated that these viruses replicated in brain microvascular endothelial cells (BMECs), which induced downregulation of tight junction proteins and increased the permeability of the barrier. Other reports demonstrated that infection of BMECs allowed the basolateral release of infectious particles, without a remarkable cytopathic effect, what might be sufficient for virus invasion. Virus replication and activation of other cells associated to the BBB, mostly astrocytes and microglia, were also reported to affect the endothelial barrier permeability. This event might occur simultaneously or after BMECs infection, being a secondary effect leading to BBB disruption. Importantly, activation of BMECs, astrocytes, and microglia by flaviviruses was associated to the expression and secretion of inflammatory mediators, which are believed to recruit leukocytes to the CNS. The leukocyte infiltrate could further mediate viral invasion through a Trojan horse mechanism and might contribute to BBB breakdown and to neurological alterations. This review discussed the previous studies regarding in vitro and in vivo models of JEV, WNV, ZIKV, YFV, DENV, and TBEV infection and addressed the pathways for BBB overcome and invasion of the CNS described for each virus infection, aiming to increment the knowledge and stimulate further discussion about the role of BBB in the neuropathogenesis of flavivirus infection.

Exacerbated inflammatory response and altered vascular function are hallmarks of dengue disease. Reactive oxygen species (ROS) production has been associated to endothelial barrier disturbance and microvascular alteration in distinct pathological conditions. Increased ROS has been reported in in vitro models of dengue virus (DENV) infection, but its impact for endothelial cell physiology had not been fully investigated. Our group had previously demonstrated that infection of human brain microvascular endothelial cells (HBMEC) with DENV results in the activation of RNA sensors and production of proinflammatory cytokines, which culminate in cell death and endothelial permeability. Here, we evaluated the role of mitochondrial function and NADPH oxidase (NOX) activation for ROS generation in HBMEC infected by DENV and investigated whether altered cellular physiology could be a consequence of virus-induced oxidative stress. DENV-infected HBMECs showed a decrease in the maximal respiratory capacity and altered membrane potential, indicating functional mitochondrial alteration, what might be related to mtROS production. Indeed, mtROS was detected at later time points after infection. Specific inhibition of mtROS diminished virus replication, cell death, and endothelial permeability, but did not affect cytokine production. On the other hand, inhibition of NOX-associated ROS production decreased virus replication and cell death, as well as the secretion of inflammatory cytokines, including IL-6, IL-8, and CCL5. These results demonstrated that DENV replication in endothelial cells induces ROS production by different pathways, which impacts biological functions that might be relevant for dengue pathogenesis. Those data also indicate oxidative stress events as relevant therapeutical targets to avoid vascular permeability, inflammation, and neuroinvasion during DENV infection.