UC Berkeley

Dengue virus (DENV) is the most medically-important arbovirus worldwide, and infection with any of the four serotypes of the virus (DENV1-4) can lead to a range of outcomes, from inapparent infection to classical dengue fever (DF) to dengue hemorrhagic fever/dengue shock syndrome (DHF/DSS), severe disease manifestations characterized by endothelial barrier dysfunction and vascular leak. Non-structural protein 1 (NS1) is the only protein secreted from DENV-infected cells and is produced as a monomer, is found in association with cell membranes as a dimer, and is secreted as a hexamer. NS1 is a key component of the viral replication complex and has been implicated in immune evasion and pathogenesis. Recently, we and others described novel roles for NS1 in directly triggering endothelial barrier dysfunction and inducing inflammatory cytokine production from human immune cells, contributing to vascular leak in vivo. In this dissertation, the mechanisms behind NS1-induced pathogenesis of the endothelium are explored.

Using in vitro techniques, we found that NS1 from all four DENV serotypes can induce endothelial hyperpermeability of human pulmonary microvascular endothelial cells (HPMEC) and showed that this effect is mediated by endothelial sialidases, cathepsin L, and heparanase. These enzymes contribute to degradation of the endothelial glycocalyx-like layer (EGL), an important determinant of endothelial barrier function, and inhibition of sialidases, cathepsin L, and heparanase prevented NS1-induced endothelial hyperpermeability. Next, we found that NS1 does not trigger the production of inflammatory cytokines from human dermal microvascular endothelial cells (HMEC-1) in vitro and showed that neither TLR4 nor TNF-α signaling contribute to NS1-mediated vascular leak in the dermis of mice in vivo. Further, we demonstrated that inhibition of sialidases, cathepsin L, and heparanase prevents NS1-induced hyperpermeability and EGL degradation of HMEC-1 in vitro and vascular leak in the mouse dermis in vivo. We then sought to elucidate the early events that occur immediately following NS1 binding to endothelial cells. Using a glycosylation mutant of NS1, N207Q, we found that the N207 glycosylation site is not required for binding of NS1 to HPMEC but is required for NS1-induced hyperpermeability and EGL degradation. Further, we observed that wild-type NS1 is internalized by endothelial cells via clathrin-mediated endocytosis and is trafficked to the endosome. While the N207Q mutant is internalized, it is not as efficient as wild-type, is not dependent on clathrin, and does not lead to endosomal localization.

We then broadened our studies to include NS1 proteins from other flaviviruses, including Zika (ZIKV), West Nile (WNV), Japanese encephalitis (JEV), and yellow fever (YFV) viruses. We found that these proteins selectively bind to and alter the permeability of human endothelial cell monolayers from the lung, dermis, umbilical vein, brain, and liver in vitro and, remarkably, cause vascular leakage upon inoculation into mice in a tissue-dependent manner, reflecting the pathophysiology of each flavivirus. Mechanistically, each flavivirus NS1 protein leads to differential disruption of key endothelial glycocalyx components after activation of sialidases, cathepsin L, and heparanase.

Finally, we sought to determine whether sera from individuals vaccinated with Takeda’s live-attenuated Tetravalent Dengue Vaccine candidate (TDV) could protect against NS1-induced hyperpermeability of HPMEC in vitro. We found that sera from DENV-naïve individuals at day 0 pre-vaccination did not protect against DENV2 NS1-induced hyperpermeability, while day 0 sera from DENV-pre-immune subjects provided varying levels of protection. However, all day 120 post-vaccination samples from both DENV-naïve and pre-immune subjects abrogated DENV2 NS1-induced endothelial hyperpermeability, and, across all serum samples, the magnitude of protection correlated with the respective anti-NS1 antibody concentration. We observed a similar pattern when evaluating protection against DENV2 NS1-induced EGL degradation. Lastly, we found that serum from vaccinees could cross-protect against NS1 from DENV1, DENV3, and DENV4, and this protection also correlated with anti-NS1 antibody concentration.

Taken together, the work included in this dissertation represents a substantial advancement of the field’s understanding of NS1 and its pathogenesis and further supports a key role for NS1 in dengue disease. Although a number of novel findings are described here, additional important questions remain unanswered, specifically regarding the molecular pathways and players involved in NS1-mediated pathogenesis as well as broader topics such as the relative contributions of NS1 in the greater context of the pathogenesis of dengue and other flaviviral diseases. Nonetheless, this work provides considerable mechanistic insight into NS1 pathogenesis and establishes NS1 as a key target for therapeutic intervention and important component of dengue vaccines.