The phenotype of endothelial cells (EC) in different regions of the circulation and functional responses of EC to stimuli, may be conditioned by local haemodynamics and undergo changes if flow conditions are modified. Different levels or patterns of shear stress applied to endothelial cells regulate inflammatory responses (such as leukocyte recruitment) and angiogenic responses (such as motility). The association between oscillatory, low time-average shear stress and the development of atheroma is well known, but the level of shear stress also more generally modulates responses to cytokines. In mice or rats, increases in shear stress in the microcirculation lead to capillary growth, and various studies indicate effects of shear stress on related endothelial cellular responses such as migration or proliferation. When shear stress is changed, rapid responses at the cellular level include increased nitric oxide production, and then gene expression is modified over hours to days. The signalling basis for the different responses is uncertain, but mechanotransduction has variously been demonstrated to operate through receptor tyrosine kinases, integrin-substrate interactions, inter-cellular junctional molecules such as CD31 and VE-cadherin, and changes in permeability of membrane ion-channels. To link signalling mechanisms to changes in functional responses at different levels of shear stress, we developed models in which human or murine EC of various types were cultured in glass capillaries coated with desired substrates. These constructs were conditioned by different levels of shear stress for different periods, or exposed to changes in shear. Conditioning was combined with treatment with cytokines such as tumour necrosis factor-α (TNF) and interleukin-1β (IL-1), and adhesion and migration of flowing neutrophils analysed as an ‘inflammatory’ readout. Alternatively, closure of ‘wounds’ made in endothelial monolayers was followed over time, as a measure of motility. Under all conditions, cells could be extracted and their gene expression analysed through quantitative RT-PCR. These studies were combined with use of siRNA to reduce expression of the putative mechanotransducers β1- or β3-integrins, or CD31, and compared with use of EC from mice lacking CD31 expression. Initial studies showed that conditioning of several types of human EC (HUVEC from umbilical veins; HUAEC from umbilical arteries; HCAEC from coronary arteries) for 24h at increasing shear stress powerfully suppressed responses to TNF, e.g., judged by neutrophil recruitment (Sheikh et al., 2003; unpublished data). This shear-induced reduction in sensitivity to TNF was less effective when expression of β1-integrin or of CD31 was suppressed before application of shear. Suppression of β3-integrin expression had no effect on responses. In the wound healing assay, exposure of human or murine EC to shear stress tended to align their migration with the direction of flow, but this resulted in a net decrease in the rate at which the gap closed. Use of siRNA against CD31 in the HUVEC, or testing of EC from mice lacking CD31, indicated that expression of CD31 was not required for the shear-induced effects in this model. Furthermore, while a panel of genes linked to inflammation and angiogenesis were shown to be shear-sensitive in their expression, this shear-sensitivity was not uniformly modified by reduction of expression of CD31 or β1-integrin. Thus, while these adhesion molecules are implicated in shear-induced signalling to modify function of EC, different functional and genomic responses to shear may not necessarily be mediated through the same pathway. Taken together, these studies suggest that local shear-conditioning of endothelial cells contributes to vessel-, organ- and stimulus-specificity in inflammatory and angiogenic responses. This specificity may play a role in predisposition of certain sites to development of disease, but also offers insight into routes by which these important pathophysiological responses might be modulated for therapeutic benefit.