Angiopoietin-2-induced blood-brain barrier compromise and increased stroke size are rescued by VE-PTP-dependent restoration of Tie2 signaling

Abstract

The homeostasis of the central nervous system is maintained by the blood-brain barrier (BBB). Angiopoietins (Ang-1/Ang-2) act as antagonizing molecules to regulate angiogenesis, vascular stability, vascular permeability and lymphatic integrity. However, the precise role of angiopoietin/Tie2 signaling at the BBB remains unclear. We investigated the influence of Ang-2 on BBB permeability in wild-type and gain-of-function (GOF) mice and demonstrated an increase in permeability by Ang-2, both in vitro and in vivo. Expression analysis of brain endothelial cells from Ang-2 GOF mice showed a downregulation of tight/adherens junction molecules and increased caveolin-1, a vesicular permeability-related molecule. Immunohistochemistry revealed reduced pericyte coverage in Ang-2 GOF mice that was supported by electron microscopy analyses, which demonstrated defective intra-endothelial junctions with increased vesicles and decreased/disrupted glycocalyx. These results demonstrate that Ang-2 mediates permeability via paracellular and transcellular routes. In patients suffering from stroke, a cerebrovascular disorder associated with BBB disruption, Ang-2 levels were upregulated. In mice, Ang-2 GOF resulted in increased infarct sizes and vessel permeability upon experimental stroke, implicating a role of Ang-2 in stroke pathophysiology. Increased permeability and stroke size were rescued by activation of Tie2 signaling using a vascular endothelial protein tyrosine phosphatase inhibitor and were independent of VE-cadherin phosphorylation. We thus identified Ang-2 as an endothelial cell-derived regulator of BBB permeability. We postulate that novel therapeutics targeting Tie2 signaling could be of potential use for opening the BBB for increased CNS drug delivery or tighten it in neurological disorders associated with cerebrovascular leakage and brain edema.

Keywords: Angiopoietin-2; Blood–brain barrier (BBB); Endothelium; Permeability; Stroke; Tie2 signaling; Vascular endothelial protein tyrosine phosphatase (VE-PTP).

Fig. 1

Ang-2 increases brain endothelial permeability in vitro. a MBMECs from WT and Ang-2 GOF brains were seeded on transwell inserts and transferred to a cellZscope^® system to obtain continuous TEER values. The inset shows a transwell insert along with top and bottom electrodes typical of a cellZscope^® device. The graph shows a representative experiment that indicates reduced TEER values in Ang-2 GOF MBMECs compared to WT that is sustained up to 48 h post-seeding when the monolayers reach confluency. b GOF MBMECs showed lower TEER than control cells at 24 and 48 h. c WT-MBMECs treated with hAng-2 resulted in reduced TEER. d hAng-2 treatment resulted in increased tracer flux across MBMECs (b–d, n = 3)

Fig. 2

Permeability analysis of Ang-2 GOF mice in vivo. a Permeability to LY and Texas Red-3 kD/TMR-dextran was higher in GOF brains (LY: n = 7, TXR: WT n = 12; GOF n = 16). b Evans blue dye permeability was not altered (n = 6). c Increased permeability of 3 kD TMR-dextran in Ang-2 GOF mice in cortex and subcortical white matter (SWM) (n = 3, scale bars 10 µm)

Fig. 3

Immunohistochemistry analysis of pericytes in Ang-2 GOF mice. a Pericytes (desmin+) were decreased in GOF mice utilizing 10 µm cryosections for analysis (n = 8, scale bars 25 µm). b 50 µm vibratome sections of Ang-2 GOF and WT mice revealed decrease in pericyte area/vessel area [%] (n = 3, scale bars 10 µm) indicating decreased pericyte coverage of the vessels in GOF mice

Fig. 4

Ultrastructural analysis of Ang-2 GOF mice. a EM revealed swollen astrocytic endfeet, degenerating pericytes and more vesicles and gaps (arrows) between ECs in GOF. AEF astrocytic endfeet, BL basal lamina, PC pericyte (scale bar sizes are indicated in the images). b The pericyte coverage analyzed from EM images is decreased in Ang-2 GOF mice (100 vessels were analyzed; n = 5). c Increased pericyte degeneration was observed in Ang-2 GOF mice (100 vessels analyzed; n = 5). d The gaps within EC junction increased in Ang-2 GOF mice (100 vessels analyzed; n = 5). e The number of cortical vesicles of ECs in Ang-2 GOF mice increased (20 vessels of similar size 4–5 µm lumen per mouse were analyzed; n = 3)

Fig. 5

Detection of the glycocalyx and plasma tracers by EM analysis: a lanthanum nitrate was used to detect the glycocalyx (black arrows) in mouse brain vessels. Glycocalyx was decreased/disrupted from ~300 to ~100 nm in GOF mice. b HRP was intravenously injected into WT and Ang-2 GOF mice and circulated for 30 min. Representative images of Ang-2 GOF mice revealed HRP-vesicles (black arrows) with affected endothelial junctions (red arrows). Quantitative analysis indicated more HRP-positive vesicles in Ang-2 GOF mice (20 vessels analyzed; n = 3; scale bar sizes are indicated in the images). c MBMV Western blots showed decreased VE-cadherin and claudin-5 in GOF, whereas caveolin-1 was upregulated (n = 3)

Fig. 6

Ang-2 expression analysis in normal brain and human stroke samples (grade I–III). a, b H&E/IHC staining of human stroke samples showed higher Ang-2 expression in the stroke area (n = 13). NAWM normal appearing white matter, NAGM normal appearing gray matter, i infarct, pn penumbra. c Serum ELISA from stroke (n = 4-lacunar, n = 15-territorial) and healthy subjects (n = 16) showed highest Ang-2 levels in territorial strokes

Fig. 7

Analysis of stroke severity in Ang-2 GOF mice. a pMCAO infarcts were larger in GOF (24 h, WT, GOF n = 4; 72 h, WT n = 10, GOF n = 7; 7 days, WT, GOF n = 5). b IgG permeability in the infarct area was higher in GOF (n = 6) compared to WT (n = 4)

Fig. 8

Effects of AKB-9785 treatment in vitro. a MBMECs pre-treated with hAng-2 increased Tie2/Akt activation upon AKB-9785 treatment (n = 3). b bEnd5 Western blots revealed Tie2/Akt activation, but a decrease (non-significant) in pVE-cadherin by AKB-9785 (n = 3). c MBMECs pre-treated with recombinant hAng-2 [500 ng/ml] followed by AKB-9785 [1 µM] treatment revealed increased TEER values. d The in vitro permeability tracer experiment with a 3 kD Texas Red-dextran showed a decrease in permeability with AKB-9785 + hAng-2 compared to hAng-2 treatment alone. This difference was not detected in the presence of filipin. e The phosphorylation of Akt in Western Blot analysis could only be identified in MBMECs, but not pericytes (n = 3)

Fig. 9

Therapeutic targeting of Tie2 signaling using AKB-9785 in WT mice subjected to tMCAO. a TTC staining after tMCAO (3/group) showed significantly smaller infarcts 24 h post-stroke upon AKB-9785 treatment (n = 13). b An mNSS behavioral analysis of control and mice treated with AKB-9785 (30 mg/kg) showed a neurological improvement in treated mice 72 h after stroke incidence (24 h: control n = 8, AKB-9785 n = 9; 72 h: control n = 7, AKB-9785 n = 9). The neurological deficit scoring is described in “Materials and methods” and is composed of testing flexion, gait, coordination and sensory functions. The total scoring included all four behavioral tests (2-tailed unpaired non-parametric Mann–Whitney test). c IgG staining showed significantly lower permeability in AKB-9785 treated mice (n = 6). d Western blots of stroke hemisphere showed increased Akt activation in AKB-9785-treated mice (n = 7). e Tie2/pTie2 staining of the stroke hemisphere revealed significantly higher pTie2, but not total Tie2 levels in AKB-9785-treated animals (n = 5)