Disturbed flow increases endothelial inflammation and permeability via a Frizzled-4-β-catenin-dependent pathway

ABSTRACT Multidirectional or disturbed flow promotes endothelial dysfunction and is associated with early atherogenesis. Here we investigated the role of Wnt signalling in flow-mediated endothelial dysfunction. The expression of Frizzled-4 was higher in cultured human aortic endothelial cells (ECs) exposed to disturbed flow compared to that seen for undisturbed flow, obtained using an orbital shaker. Increased expression was also detected in regions of the porcine aortic arch exposed to disturbed flow. The increased Frizzled-4 expression in cultured ECs was abrogated following knockdown of R-spondin-3. Disturbed flow also increased the nuclear localisation and activation of β-catenin, an effect that was dependent on Frizzled-4 and R-spondin-3. Inhibition of β-catenin using the small-molecule inhibitor iCRT5 or knockdown of Frizzled-4 or R-spondin-3 resulted in reduced expression of pro-inflammatory genes in ECs exposed to disturbed flow, as did inhibition of WNT5A signalling. Inhibition of the canonical Wnt pathway had no effect. Inhibition of β-catenin also reduced endothelial paracellular permeability; this was associated with altered junctional and focal adhesion organisation and cytoskeletal remodelling. These data suggest the presence of an atypical Frizzled-4-β-catenin pathway that promotes endothelial dysfunction in response to disturbed flow.


Fig. S2 .
Fig. S2.Efficacy and specificity of Frizzled-4 knockdown using siRNA (A-D) HAEC were exposed to flow for 48h following transfection with Frizzled-4 siRNA or scrambled controls.(A) Protein lysates were obtained from HAEC exposed to DF and the expression of Frizzled-4 assessed by western blot using calnexin as a loading control (n=6; analysis by Mann-Whitney test).(B) RNA was harvested from EC exposed to DF and the expression of Frizzled-4, Frizzled-5, Frizzled-6 and Frizzled-7 was determined by qRT-PCR using GAPDH as a housekeeping gene (n=6; analysis by Mann-Whitney test; results shown relative to DF scrambled control (dashed line)).(C-D) RNA was harvested from EC exposed to DF and UF and the expression of (C) KLF-2 and (D) NOS3 was determined by qRT-PCR using GAPDH as a housekeeping gene (n=6; analysis by two-way ANOVA with Tukey's multiple comparison test; results shown relative to DF scrambled control).

Fig. S3 .
Fig. S3.Disturbed flow increases the expression and activation of β-catenin in human aortic endothelial cells (A-B) HAEC were exposed to flow for 72h.Lysates were prepared from EC exposed to DF and UF and subject to sub-cellular fractionation.Cytosolic and nuclear fractions were analysed by western blot using (A) total β-catenin (n=6-8) and (B) active β-catenin antibodies (n=5).PDHX and TBP were used as loading controls for cytosolic and nuclear fractions respectively; analysis by Mann Whitney test; representative blots shown above.(C) HAEC were transfected with Cignal TCF/LEF reporter construct prior to flow exposure for 1h -48h.Lysates were prepared from cells exposed to DF or UF and Firefly and Renilla luciferase activity was recorded.Ratios were corrected for protein content of lysates.(C) Results shown relative to DF control at each time point and analysed by Mann-Whitney test at each time point (n=5).(D) Results shown relative to DF 1h and analysed by one-way ANOVA (n=5).

Fig. S4 .
Fig. S4.Expression of pro-inflammatory genes in EC exposed to disturbed flow following inhibition or knockdown of β-catenin (A) HAEC were exposed to flow for 72h with iCRT5 (50 µM) added for the duration of flow exposure (72h).RNA was harvested from EC exposed to DF and the expression of E-SEL, MCP-1 and VCAM1 determined by qRT-PCR using GAPDH as a housekeeping gene (n=5-8; analysis by Mann-Whitney test).(B-C) HAEC were exposed to flow for 72h following transfection with β-catenin siRNA (siβ-cat).(B) The expression of β-catenin protein was assessed by western blot using PDHX as a loading control (n=6; analysis by analysis by unpaired t-test; representative blot shown next to the graph).(C) RNA was harvested from EC exposed to DF and the expression of pro-inflammatory genes assessed by qRT-PCR using GAPDH as a housekeeping gene (n=6; analysis by Mann-Whitney test).

Fig. S5 .
Fig. S5.Permeability and morphology of HAEC(A) HAEC were cultured on biotinylated-gelatin and exposed to flow for 72h then treated with iCRT5 (50µM) for the last 24h of flow exposure.FITC-avidin was added to monolayers immediately after flow cessation.Images show areas where FITC-avidin binds to biotinylatedgelatin underlying EC.Cells were counterstained with an anti-VE-cadherin antibody (images shown are maximum projections of z-stacks; scale = 50µm).(B) HAEC were fixed and stained with VE-cadherin and DRAQ5 and viewed at x10 magnification.The length-to-width ratio of cells in at least 3 fields of view per experiment (approx.200 cells per field) was determined in cells exposed to DF and treated with iCRT5 (50 µM), IWR-1 (10 µM) or DKK-1 (250 ng.ml -1 ).The length-to-width ratio of EC exposed to UF are shown for comparison (n=3; analysis by Kruskal-Wallis test with uncorrected Dunn's test).(C) HAEC were exposed to flow for 72h and treated with iCRT5 (50µM) for the last 24h of flow exposure.HAEC were fixed and stained with ZO-1 and DRAQ5.Mean fluorescence intensity of standardised regions of interest spanning cell junctions was used to quantify ZO-1 expression (n=4; analysis by Mann-Whitney test) (D) HAEC were exposed to flow for 72h and treated with iCRT5 (50µM) for the last 24h of flow exposure or (E) transfected with Frizzled-4 siRNA (siFzd4) and exposed to flow for 48h.RNA was harvested from EC exposed to DF and expression of TIE2, α-smooth muscle actin (α-SMA) and SNAIL assessed by qRT-PCR using GAPDH as a housekeeping gene (n=5-6; analysis by Mann-Whitney test).

Fig
Fig. S6.SFRP-1 reduces inflammatory signalling in EC exposed to disturbed flow and alters cytoskeletal organisation (A-C) HAEC were exposed to flow for 72h and treated with SFRP-1 (200 ng.ml -1 ) for the last 24h of flow exposure.(A) RNA was harvested from EC exposed to DF and the expression of pro-inflammatory genes assessed by qRT-PCR using GAPDH as a housekeeping gene (n=3-4; analysis by Mann-Whitney test).(B-C) EC were fixed and incubated with anti-VE-cadherin antibody, 488-phalloidin or anti-ZO-1 antibody and DRAQ5 nuclear stain (n=4; scale = 50 µm; representative images shown).

Fig. S8 .
Fig. S8.Expression and phosphorylation of JNK in EC exposed to flowHAEC were exposed to flow for 72h and treated with iCRT5 (50µM) for the last 24h of flow exposure.Lysates from EC exposed to DF and UF were assessed by western blot using phospho-JNK and total JNK antibodies.Calnexin was used as a loading control.Results analysed by one-way ANOVA (n=5; representative blots shown above).