Here, we discuss pathophysiological approaches to the defining of endothelial dysfunction criteria (i.e., endothelial activation, impaired endothelial mechanotransduction, endothelial-to-mesenchymal transition, reduced nitric oxide release, compromised endothelial integrity, and loss of anti-thrombogenic properties) in different in vitro and in vivo models. The canonical definition of endothelial dysfunction includes insufficient production of vasodilators, pro-thrombotic and pro-inflammatory activation of endothelial cells, and pathologically increased endothelial permeability. Among the clinical consequences of endothelial dysfunction are arterial hypertension, macro- and microangiopathy, and microalbuminuria. We propose to extend the definition of endothelial dysfunction by adding altered endothelial mechanotransduction and endothelial-to-mesenchymal transition to its criteria. Albeit interleukin-6, interleukin-8, and MCP-1/CCL2 dictate the pathogenic paracrine effects of dysfunctional endothelial cells and are therefore reliable endothelial dysfunction biomarkers in vitro, they are non-specific for endothelial cells and cannot be used for the diagnostics of endothelial dysfunction in vivo. Conceptual improvements in the existing methods to model endothelial dysfunction, specifically, in relation to the blood–brain barrier, include endothelial cell culturing under pulsatile flow, collagen IV coating of flow chambers, and endothelial lysate collection from the blood vessels of laboratory animals in situ for the subsequent gene and protein expression profiling. Combined with the simulation of paracrine effects by using conditioned medium from dysfunctional endothelial cells, these flow-sensitive models have a high physiological relevance, bringing the experimental conditions to the physiological scenario.