Syndecan 4 is required for endothelial alignment in flow and atheroprotective signaling

Abstract

Atherosclerotic plaque localization correlates with regions of disturbed flow in which endothelial cells (ECs) align poorly, whereas sustained laminar flow correlates with cell alignment in the direction of flow and resistance to atherosclerosis. We now report that in hypercholesterolemic mice, deletion of syndecan 4 (S4(-/-)) drastically increased atherosclerotic plaque burden with the appearance of plaque in normally resistant locations. Strikingly, ECs from the thoracic aortas of S4(-/-) mice were poorly aligned in the direction of the flow. Depletion of S4 in human umbilical vein endothelial cells (HUVECs) using shRNA also inhibited flow-induced alignment in vitro, which was rescued by re-expression of S4. This effect was highly specific, as flow activation of VEGF receptor 2 and NF-κB was normal. S4-depleted ECs aligned in cyclic stretch and even elongated under flow, although nondirectionally. EC alignment was previously found to have a causal role in modulating activation of inflammatory versus antiinflammatory pathways by flow. Consistent with these results, S4-depleted HUVECs in long-term laminar flow showed increased activation of proinflammatory NF-κB and decreased induction of antiinflammatory kruppel-like factor (KLF) 2 and KLF4. Thus, S4 plays a critical role in sensing flow direction to promote cell alignment and inhibit atherosclerosis.

Keywords: atherosclerosis; mechanotransduction; polarity; shear stress.

Fig. 1.

Effect of S4 deletion on atherosclerosis. (A) Thoracic aortas from male DKO hypercholesterolemic mice with or without S4 (S4^+/+ or S4^−/−) after 16 or 20 wk on PD. Aorta were stained with Oil red O to reveal the atherosclerotic lesions (in white/red). (B) Quantification of the lesions area, normalized to the total area of the thoracic aorta (n = 6 for DKO/S4^+/+ and n = 7 for DKO/S4^−/−; P < 0.01). (C) Detailed view of the intercostal arteries (white stars) after 20 wk of PD.

Fig. 2.

Endothelial morphology in the thoracic aorta. En face view of the thoracic aorta from wild-type or S4^−/− mice. Tissue was stained for beta-catenin to mark cell borders (A) or with phalloidin and DAPI to mark stress fibers and nuclei (B). Stress fiber alignment: stress fibers were considered “aligned” if their direction was parallel to the direction of the cell major axis (within 30° of the major axis) and “misaligned” if not (n < 300 cells, four arteries per condition).

Fig. 3.

S4 is required for endothelial cell polarity in flow. (A) HUVECs stably expressing scrambled shRNA or S4 shRNA (#119) were untreated (no flow) or subjected to laminar FSS at 12 dynes/cm² for 16 h [(B) quantification, n > 3,000 cells/experiment from 15 independent experiments] or 10% uniaxial cyclic stretch (1 Hz) for 4 h [(C) quantification, n > 2,000 cells/experiment from seven independent experiments]. Cells were fixed and labeled with DAPI (red) and phalloidin (cyan). (D) Rescue of S4 knockdown (shRNA #121) by adenoviral reexpression of rat S4. Western blot of S4 and actin as a loading control; all samples were sheared for 16 h. (E) Nuclear orientation was quantified to characterize cell orientation relative to flow direction (n > 3,000 cells, from at least four independent experiments). Values are means ± SEM.

Fig. 4.

Effect of S4 knock-down on fluid shear-stress responses. (A) Western blot of phosphorylated and total VEGFR2, with actin as a loading control (n = 4, mean ± SEM). (B) Average nuclear eccentricity for all the nuclei in each experiment was quantified using the Matlab Eccentricity function (n ≥ 4 individual experiments, n > 3,000 nuclei per experiment; NF, no flow). (C) To assess NF-κB translocation, cells were stained for p65, imaged, and analyzed as described in Methods. The translocation factor is the total nuclear p65 fluorescence relative to the total cytosolic p65 fluorescence. Results for each experiment were then normalized to the static scramble cells. At least 2,000 cells were analyzed per condition, in at least three independent experiments (*P < 0.05, values are mean ± SEM). (D) S4, KLF2, and KLF4 message levels, relative to the static scramble mRNA for each individual experiment. mRNA levels were normalized to GAPDH (n = 4; ***P < 0.001 versus static scramble; ###P < 0.001 versus flow scrambled). SCR, scrambled shRNA; S4 #1, S4 shRNA #119; S4 #2, S4 shRNA #121. Values are mean ± SEM.