Vasoactive Soluble Endoglin: A Novel Biomarker Indicative of Reperfusion after Cerebral Large-Vessel Occlusion

Abstract

Now that mechanical thrombectomy has substantially improved outcomes after large-vessel occlusion stroke in up to every second patient, futile reperfusion wherein successful recanalization is not followed by a favorable outcome is moving into focus. Unfortunately, blood-based biomarkers, which identify critical stages of hemodynamically compromised yet reperfused tissue, are lacking. We recently reported that hypoxia induces the expression of endoglin, a TGF-β co-receptor, in human brain endothelium in vitro. Subsequent reoxygenation resulted in shedding. Our cell model suggests that soluble endoglin compromises the brain endothelial barrier function. To evaluate soluble endoglin as a potential biomarker of reperfusion (-injury) we analyzed its concentration in 148 blood samples of patients with acute stroke due to large-vessel occlusion. In line with our in vitro data, systemic soluble endoglin concentrations were significantly higher in patients with successful recanalization, whereas hypoxia alone did not induce local endoglin shedding, as analyzed by intra-arterial samples from hypoxic vasculature. In patients with reperfusion, higher concentrations of soluble endoglin additionally indicated larger infarct volumes at admission. In summary, we give translational evidence that the sequence of hypoxia and subsequent reoxygenation triggers the release of vasoactive soluble endoglin in large-vessel occlusion stroke and can serve as a biomarker for severe ischemia with ensuing recanalization/reperfusion.

Keywords: biomarker; brain endothelium; endoglin; hypoxia; mechanical thrombectomy; reperfusion injury; shedding; stroke.

Figure 1

Concentration of sENG in venous plasma samples of stroke patients with LVO in the anterior circulation. Blood samples were drawn 18.2 (±8.4) hours after MT and subjected to ELISA. In the 129 patients undergoing MT, sENG levels were significantly higher after complete (TICI 3) recanalization (TICI 0–1: n = 10, 20% received rt-PA; TICI 2: n = 94, 38.5% received rt-PA; TICI 3: n = 25, and 60% received rt-PA; there was no significant difference in the ASPECT scores between these groups) (A). In patients with successful recanalization, larger infarct volumes, as reflected by a lower ASPECT score at admission, resulted in increased sENG concentrations (148 patients including 19 without MT) (B). Application of rt-PA alone or in addition to MT did not influence shedding of sENG (black bar reflects 60 patients receiving rt-PA (80% with MT); white bar shows 88 patients without rt-PA treatment (92% with MT) (C)). Graphs depict means and standard deviation. Non-parametric Kruskal–Wallis test followed by Dunn’s post-test in A, T-test by the nonparametric Kolmogorov–Smirnov test in B and C; (* p < 0.05, ** p < 0.01).

Figure 2

Positive correlation of sENG concentrations with the NIHSS 72 h after stroke onset in patients with partial or complete (corresponding to TICI 2C or TICI 3) recanalization. Spearman correlation coefficient (n = 39, r = 0.44, p = 0.0047).