Flow-Mediated Susceptibility and Molecular Response of Cerebral Endothelia to SARS-CoV-2 Infection

Abstract

Background and purpose: Severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) infection is associated with an increased rate of cerebrovascular events including ischemic stroke and intracerebral hemorrhage. The mechanisms underlying cerebral endothelial susceptibility and response to SARS-CoV-2 are unknown yet critical to understanding the association of SARS-CoV-2 infection with cerebrovascular events.

Methods: Endothelial cells were isolated from human brain and analyzed by RNA sequencing. Human umbilical vein and human brain microvascular cells were used in both monolayer culture and endothelialized within a 3-dimensional printed vascular model of the middle cerebral artery. Gene expression levels were measured by quantitative polymerase chain reaction and direct RNA hybridization. Recombinant SARS-CoV-2 S protein and S protein-containing liposomes were used to measure endothelial binding by immunocytochemistry.

Results: ACE2 (angiotensin-converting enzyme-2) mRNA levels were low in human brain and monolayer endothelial cell culture. Within the 3-dimensional printed vascular model, ACE2 gene expression and protein levels were progressively increased by vessel size and flow rates. SARS-CoV-2 S protein-containing liposomes were detected in human umbilical vein endothelial cells and human brain microvascular endothelial cells in 3-dimensional middle cerebral artery models but not in monolayer culture consistent with flow dependency of ACE2 expression. Binding of SARS-CoV-2 S protein triggered 83 unique genes in human brain endothelial cells including upregulation of complement component C3.

Conclusions: Brain endothelial cells are susceptible to direct SARS-CoV-2 infection through flow-dependent expression of ACE2. Viral S protein binding triggers a unique gene expression profile in brain endothelia that may explain the association of SARS-CoV-2 infection with cerebrovascular events.

Keywords: cerebrovascular circulation; endothelial cells; endothelium, vascular; models, theoretical; viruses.

Figures

Figure 1.. Flow-dependent ACE2 expression in cerebral endothelial cells.

Average FPKM of gene expression derived from RNA-sequencing of normal white matter (NWM; gray) and normal CD31+ endothelial cells (NEC; orange) (n=2) (A). Fold-expression by qPCR for human ACE2 (hACE2) relative to GAPDH in HUVEC cells by vessel size and shear stress (15 dynes/cm2) (p<0.0001 by ANOVA; B). Fold-expression levels of HUVECs and HBMECs low shear stress vs. high shear stress perfusion culture (p<0.0001 by two-way ANOVA; adjusted p-values provided Tukey’s post-hoc; C). Immunocytochemistry for ACE2 in monolayer HBMECs and HUVECs (D). Error bars represent S.D. Scale bars = 100 μm.

Figure 2.. Regional variation in ACE2 levels using a 3D-printed MCA model.

Workflow to generate 3D-printed MCA stenosis model (A). Representative immunocytochemistry for ACE2 in monolayer HBMECs as well as proximal, stenotic, and distal segments of a human brain microvascular endothelialized middle cerebral artery 3D model with computational fluid dynamic modeling of wall shear stress (WSS) across the stenotic model (B). Normalized ACE2 levels (ACE2/Phalloidin ratio) in monolayer and segments of MCA 3D model (p=0.043 by one-way ANOVA; adjusted p=0.046 by Tukey’s post-hoc comparison of 2D vs. 3D culture; C). Scale bars = 100 μm.

Figure 3.. Flow-dependent susceptibility of brain endothelia to Sars-CoV-2 Spike Protein liposomes.

Liposomes coated with S-protein were created as described, incubated with soluble hACE2 (0–5 μg) and S-protein:ACE2 binding detected by ACE2:S-protein sandwich ELISA (left) and intrinsic rhodamine signal (right). Monolayer HBMECs were exposed to control (upper) or S-protein coated liposomes (lower) (1:100) for 24 hrs followed by immunocytochemistry for hACE2 (purple), phalloidin (green) and liposome bound rhodamine (white) (B). Maximum intensity projection of 3D MCA stenosis model endothelialized with HBMECs after 24 hr exposure to control (upper) or S-protein coated liposomes (lower) (1:500; 0.2 μg) during perfusion culture. Bound S-protein liposomes were detected within multiple luminal HBMECs (arrow, C). Scale bars = 100 μm.

Figure 4.. Brain endothelial-specific differential gene expression triggered by recombinant Sars-CoV-2 Spike Protein exposure.

Normalized differential gene expression measured by direct RNA hybridization after 24 hr exposure of HBMECs to 5 μg/mL Sars-CoV-2 Spike Protein trimer (S-protein) (A). Differential gene expression in HBMEC after S-protein exposure normalized to expression levels in HUVECs (B). Dashed line represents -log adjusted p-value. LogFC = log fold change.