Shear-induced Endothelial NOS Activation and Remodeling via Heparan Sulfate, Glypican-1, and Syndecan-1

Enzymatic Removal of HS

Recently, we demonstrated that cultured BAEC contain abundant HS embedded within a substantial glycocalyx ⁴⁰. Using immunofluorescnce, we verified the presence of HS on untreated confluent BAEC monolayers and confirmed the efficiency of HepIII induced HS cleavage. Approximately 67% of the HS surface was fluorescently labeled with HS antibodies under normal conidtions, whereas treatment of the BAEC monolayer with 15 mU/mL of HepIII for 2 hours reduced coverage to 46% (Figure 1).

Knockdown of GPC1 and SDC1

Densitometric analysis of Western blots demonstrated that transfecting BAEC with glypican-1 or syndecan-1 shRNA successfully knocked down expression of the targeted glycocalyx core proteins. Transfection of glypican-1 shRNA resulted in a 72.5 ± 0.05% reduction of glypican-1 (Figure 2C). However, glypican-1 shRNA also reduced non-target syndecan-1 protein expression by 36.7 ± 0.03% (Figure 2C). Syndecan-1 shRNA resulted in 62.0 ± 0.09% reduction of syndecan-1 in BAEC (Figure 2D). Syndecan-1 shRNA had no effect on non-target glypican-1 expression (Figure 2D). For control purposes, we initially confirmed that BAEC treatment with the transfection media, reagent, and vector—OptiMEM, Optifect, and scrambled shRNA or GFP-expressing vector—induced no toxicity and no change in SDC1 or GPC1 expression levels (Figure 2B). We visualized the GFP-expressing vector transfected in live BAEC to see in real-time how effectively the RNA interfering shRNA was delivered into and then stabilized in BAEC cultures. The efficiency of the stable shRNA transfection, as indicated by the percentage of BAEC expressing the GFP control shRNA, was 80% to 90% (Figure 2A). HS immunofluorescence results from additional control experiments confirmed the loss of HS coincident with primary reduction of GPC1 core protein expression and secondary reduction of SDC1 core protein expression using the GPC1 shRNA (Supplemental Figure 1). These control experiments were not performed on SDC1 shRNA treated BAEC, due to the untimely discontinuation of the commercially available HS antibody that was used for this study.

HS and GPC1 Knockdown, but not SDC1, Inhibit Flow-induced eNOS Activation

Shown in figure 3 are representative Western blots (Figs 3A, 3C, and 3E) and their densitometric quantification (Figs 3B, 3D, and 3F) from BAEC exposed to (i) no-flow control and (ii) 3-hours shear stress. Figures 3A and 3B correspond to HS-degraded monolayers, figures 3C and 3D to GPC1-silenced monolayers and figures 3E and 3F to SDC1-silenced monolayers. In addition to probing for ser1177 phosphorylated eNOS (p-eNOS), we probed for total eNOS and β-actin. Total eNOS was insensitive to shear stress and other treatments, as was β-actin, the internal loading control. The densitometric analysis shows that, the expression of p-eNOS (normalized by total eNOS) in BAEC exposed to shear stress was significantly up regulated in the 3 sets of experiments by 2.1 ± 0.5 to 2.9 ± 0.9 fold relative to unsheared controls.

Up regulation of p-eNOS by shear stress was completely abolished by HepIII treatment (Figs 3A and 3B). Notably, treating unsheared control BAEC with HepIII had a negligible effect on p-eNOS expression compared to untreated/unsheared BAEC (Figs 3A and 3B), indicating that HS does not participate in eNOS activation in static conditions but only in the presence of flow. The shear stress- and HS-regulated changes in eNOS phosphorylation are consistent with previously reported alteration in shear-induced NO production in BAEC that was completely blocked by treatment with Hep III at the same concentration we have used [3, 4].

Knockdown of the HS-binding core protein GPC1, along with off-target low level knockdown of SDC1, completely blocked shear-induced activation of p-eNOS (Fig 3 C,D). In static controls, the p-eNOS levels of untreated BAECs and GPC1-silenced BAEC were similar. Thus, like HS, GPC1 does not participate in eNOS activation in static conditions but only in the presence of flow. The question remains whether or not SDC1 knockdown by GPC1 shRNA (Fig 2) contributes to the inhibition of shear-induced p-eNOS.

To address this point, when BAEC were transfected with SDC1 shRNA, which exclusively targets SDC1 and not GPC1 (Fig 2), flow-induced p-eNOS expression was not inhibited (Figs 3E and 3F). There was a small but significant increase in p-eNOS expression of unsheared SDC1 knockdowns compared to untreated unsheared BAEC conditions, but shear-induced p-eNOS was still up-regulated significantly relative to the SDC1 knockdown static controls.

HS and SDC1, but not GPC1, Regulate Flow-induced Endothelial Cell Remodeling

Figure 4 shows representative phase contrast images of BAEC after 24-hour exposure to static control conditions or 15 dynes/cm² average shear stress, along with quantification of cell elongation and alignment. The results indicate the significant role of HS, and the insignificant role of the HS core protein GPC1, in EC remodeling. Control endothelial cells (untreated and unsheared) exhibited a near cobblestone morphology (Fig 4A) with aspect ratios of ranging from 1.75 to 1.85, depending on the EC batch. Fold changes in aspect ratio are presented in figure 4E, so the aspect ratio of the control EC shown in figure 4A has a value of 1 (normalized). These cells show no particular orientation, as indicated by the angles of their major axis with respect to the flow direction that were distributed randomly at all angles from 0° to ± 90° (Figures 4F to 4H, grey bars). As expected, untreated ECs exposed to flow were elongated and aligned in the direction of flow after 24 hours, with a fold increase in aspect ratio to 2.1 ± 0.05 (Figures 4B and 4E). The major axes of 55 ± 5% of the flow-conditioned cells (compared to ~12% of the static conditioned cells) were oriented between 0° and ± 10° with respect to the direction of flow (Figure 4F, black bars).

HS degradation by HepIII treatment statistically significantly reduced shear- induced elongation (Fig 4 C, E) and completely abolished alignment (Fig 4 C, G). The fold increase was only 1.3 ± 0.3 in HepIII-treated sheared BAEC, compared to 2.1 ± 0.05 in untreated BAEC (Figs 4C and 4E). In comparison to untreated sheared BAEC, HepIII-treated cells were not aligned in the direction of flow with only 18 ± 3 % of cells oriented between 0° and ± 10° relative to the flow direction (Fig 4G, black bars). Knockdown of glypican-1 preserved most EC remodeling under flow. Shear-conditioned GPC1-silenced BAEC were elongated to a 2.0 ± 0.3 fold increased aspect ratio compared to 2.1 ± 0.05 in untreated BAEC (Figures 4D and 4E) and 38 ± 4 % of cells were oriented between 0° and ± 10° relative to flow (Figure 4H, black bars).

We attempted to perform similar 24-hour cell remodeling experiments with SDC1 shRNA-treated BAEC. We found, however, that BAEC monolayers with reduced SDC1 expression were unable to maintain their integrity under long term shear. Thirty to 40 percent of cells were lost from the monolayer after flow conditioning (Supplemental Figure 2). In lieu of 24-hour studies of SDC1 knockdown remodeling, 6-hour experiments were undertaken. In these experiments, we performed phase contrast microscopy together with cell aspect ratio and orientation angle analysis. In addition, confocal microscope immunofluorescence images of stained f-actin fibers were qualitatively assessed. The results of these experiments are summarized in figure 5. We observed and confirmed that control EC exhibit cobblestone shape morphology and show no preferential alignment in any angle from 0° to ± 90° angle (Figure 5A, grey e in Figure 5E, and grey columns in Figure 5F). Control BAEC f-actin fibers were organized randomly within the cell body (Figures 5A). After 6 hours of flow exposure, untreated ECs showed the beginnings of elongation with a 1.14 ± 0.02 fold increase in aspect ratio (Fig 5D). 6-hr flow-conditioned EC showed the start of inclination towards the direction of flow, with the major axes of 17 ± 3 % of the flow-conditioned cells (compared to ~8.5% of the static conditioned cells) oriented between 0° and ± 10° with respect to the direction of flow (Figure 5B and black columns in 5E). Untreated BAEC that were shear conditioned for 6 hours displayed f-actin fibers that were well oriented in the direction of the cellular long axis (Figure 5B). Knockdown of SDC1 interrupted EC remodeling during 6 hours of flow. Shear-conditioned and SDC1-silenced BAEC showed no change in aspect ratio (Fig 5D) and no preferred orientation relative to static controls (Fig 5F). Angles were distributed nearly equally in all possible directions (Figure 5C, 5D, and black columns in 5F). The f-actin fibers in sheared knockdown cells appeared just as disorganized as those of untreated and unsheared controls (Figure 5C).

Cell Culture

BAEC were obtained from Lonza Walkersville, Inc. (Walkersville, MD). Cells were grown in tissue culture dishes according to Lonza instructions in EGM-MW growth medium supplemented with 5% fetal bovine serum (FBS), 0.04% hydrocortisone, 0.4% human fibroblast growth factor, 0.1% vascular endothelial growth factor, 0.1% R3-insulin-like growth factor-1, 0.1% ascorbic acid, 0.1% human recombinant epidermal growth factor, and 0.1% gentamicin and amphotericin-B (GA-1000), all purchased from Lonza. At 70% to 90% confluence, BAEC were sub-cultured with HEPES buffered saline solution (Lonza) followed by 0.25 mg/ml trypsin/EDTA solution (Lonza) to detach cells from their substrate. The cells were transferred to fibronectin-coated glass slides (30 μg/mL, Corning), seeded at a cell density of 5,000 cells/cm², and allowed to grow to confluency for 3-5 days.

Glycocalyx Stabilization

For two hours prior to and during the shear stress experiments, BAEC were cultured in fully supplemented growth medium containing a reduced concentration of FBS (2.5%) along with bovine serum albumin (BSA; 0.5%) to stabilize the GCX. This medium is referred to as experimental medium.

HS Degradation

For short term shear stress experiments, HS was enzymatically degraded by adding F. heparinum heparinase III (HepIII) to the experimenta medium during the two hour pre-flow period. For 24-hour shear stress experiments, HS was enzymatically degraded by HepIII treatment during the initial two hour pre-flow period and also throughout the duration of the flow experiment. The purpose of maintaining the presence of HepIII enzyme for the 24-hour flow experiment was to prevent the potential onset of HS restoration that occurs hours to days after enzyme removal ⁵³. As previously described, HepIII was used at a concentration of 15 mU/mL when obtained from Sigma. Occasionally, HepIII was obtained from IBEX Technologies Inc. and the concentration used was also 15 mU/mL, determined by converting IBEX international units (I.U.) to Sigma units (1 I.U. is equivalent to 600 Sigma units). HS degradation was confirmed by immunofluorescence microscopy of aldehyde-fixed BAEC stained with anti-mouse heparan sulfate (HepSS-1) monoclonal antibody (US Biological). HS fluorescent images of untreated and HepIII-treated monolayers were recorded at using uniform camera settings. NIH ImageJ was employed to analyze the images, subtract background fluorescence, and measure fluorescence intensity by calculating the illuminated fraction (percent) of the total field of view.

GPC1 and SDC1 Silencing

GPC1 and SDC1 were gene silenced via RNA interference (RNAi). Specifically, EC monolayers were plated in 35-mm diameter culture treated petri dishes in GA-1000-free culture medium. At 60% confluence, the EC were transfected with 4 μg of bovine-specific GPC1 or SDC1 short hairpin RNA (shRNA) obtained from Santa Cruz Biotechnology. The bovine GPC1 and SDC1 shRNA plasmids were each a pool of 3 shRNA plasmids which are listed in the table 1.

12 μl of Optifect transfection reagent (Invitrogen) in 2 ml of reduced-serum OptiMEM (Invitrogen) transfection medium was used to facilitate the transfection. For control purposes, groups of EC monolayers were (i) not transfected, (ii) transfected with OptiMEM, Optifect, the vector (in which the shRNA was packaged to facilitate introduction of the shRNA in the cells), and scrambled shRNA sequences (proprietary; Santa Cruz Biotechnology), or (iii) transfected with OptiMEM, Optifect, the empty vector, and GFP. After 12 hours, the transfection medium was replaced with GA-1000-free culture medium. After another 36 hours, to generate stably transfected cell lines, 2.5 μg/ml of puromycin antibiotic was added to the culture medium along with the original 0.1% GA-1000. GPC1 and SDC1 protein suppression were demonstrated by Western blots and densitometric analyses (see further methods below). The transfection was allowed to stabilize, and cells were expanded, usually within two weeks before shear stress experiments were performed. During shear stress experiments, GPC1 and SDC1 knockdown BAEC were exposed to puromycin-containing experimental medium.

Shear Stress Experiments

BAEC were sheared in a parallel plate flow chamber (C&L Instruments) for 3 or 6 hours, or BAEC were sheared in a rotating disk flow apparatus (supplemental information) for 24 hours. In the parallel plate flow chamber, with a cell contact surface area of 6.6 cm² and a channel height of 250 μm, confluent BAEC monolayers were sheared at 15 dyne/cm². Since all cells were exposed to the same magnitude of shear stress, the entire monolayer was suitable for protein quantification. Fluid was driven through the parallel plate chamber from a multi-head peristaltic pump (Rainin) then via a vibration damper, followed by the flow chamber, and finally a receiving reservoir before being re-circulated. The rotating disk apparatus was custom designed to shear BAEC grown in a 6-well plate, in order to maximize the number of experiments that could be performed in a 24-hour time period (supplemental information). In each well, the cell contact surface was 2.7 cm² on a 3.1-cm diameter round glass coverslip fixed to the bottom of the well using sterile silicone grease (supplemental information). 6 cylindrical disks were suspended at 500 μm above 6 cell layers. Using a motor drive, the disks were simultaneously rotated at the number of rotations per minute required to generate an average shear stress of 14 dyne/cm² (supplemental information). The parallel plate and rotating disk flow systems were both maintained in an incubator at 37 °C with a supply of 5% CO₂. At the end of the experiments, samples were processed as described below.

Whole Cell Protein Extraction and Western Blotting

Whole cell protein was extracted from cells using lysis buffer containing 1 mM sodium bicarbonate, 2 mM phenylmethylsuphonyl fluoride, 1 mM sodium orthovanadate, 5 mM EDTA, and 1X complete protease inhibitor cocktail (Roche Laboratories). After scraping and collecting the cells, the extracts were sonicated briefly and frozen at −80°C overnight. Equal protein amounts (10 μg/lane) were determined by the BCA protein assay kit (Pierce) and resolved on 7.5% sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) gels. Protein was transferred to nitrocellulose membranes (Whatman), which were blocked by incubation for 30 minutes at room temperature with TBS containing 0.5% Tween-20 (TBS-T) and 5% nonfat dry milk. Membranes were then incubated overnight at 4°C with primary antibody. Primary antibodies were against Ser¹¹⁷⁷ phosphorylated eNOS (1:1000) and total eNOS (1:2500) (both from Cell Signaling), glypican-1 (1:10000; characterized and provided by Dr. Richard Margolis, NYU Langone Medical Center ⁵⁴), syndecan-1 (1:1000; clone H-174 from Santa Cruz Biotechnology), and β-actin (1:10000; Sigma). On the following day, membranes were incubated for 2 hours at room temperature with secondary label. Secondary labeling was accomplished with horseradish peroxidase-conjugated goat anti-rabbit IgG (1:10000; Santa Cruz Biotechnology) for all proteins of interest, except β-actin which was secondarily labeled with horseradish peroxidase-conjugated goat anti-mouse IgG (1:10000; Santa Cruz Biotechnology). After incubation of membranes with Immobilon Western Detection Reagents (Millipore), detection of protein bands was performed using a Carestream Image Station 4000MM Pro and Carestream Molecular Imaging (MI) software. Densitometry measurements were achieved using ImageJ.

Assessment of Cell Remodeling

To evaluate cell morphology and orientation angle, BAEC were imaged by phase contrast microscopy. ImageJ was used to trace the perimeters of individual BAEC and determine their aspect ratios (AR), defined as AR = (long axis)/(short axis). The aspect ratio was near 1 for a round or cobblestone-shaped EC and much higher than 1 for elongated cells. We also used ImageJ to calculate the angles of the BAEC long axes relative to the direction of flow. Relative angles were grouped and plotted in 20° intervals. For each condition of interest we performed 4 independent experiments. Per experiment, we examined 3 randomly distributed fields in each of two BAEC monolayers. We quantified the aspect ratios of 30 cells per field, but only when BAEC monolayers were intact.

For short term flow exposure studies in which changes in cell shape and flow-directed alignment that are less striking, cytoskeleton reorganization was also evaluated to assess cell remodeling. This was achieved by examining f-actin redistribution. BAEC were fixed in 4% paraformaldehyde for 10 minutes, permeabilized in 0.2% triton X-100, and blocked in 3% goat serum/0.2% triton X-100. Cell monolayers were then treated with Alexa Fluor 594 conjugated F-actin Phalloidin (Molecular Probes) diluted in 3% BSA. 4′,6-diamidino-2-phenylindole (DAPI) marked the EC nucleii. BAEC cytoskeleton and nucleus fluorescence were visualized with a Zeiss LSM 510 confocal laser scanning system and Zeiss LSM software, a 60X/Oil objective, and by z-stacking.

Statistical Analysis

HS fluorescence intensity data sets were averaged and normalized by untreated control values. Results are expressed as mean ± SEM. The data set was analyzed using the unpaired two-tailed t-test to determine the statistical significance of mean differences. SDC1 and GPC1 protein levels were averaged, normalized by baseline values, expressed as mean ± SEM, and statistically analyzed using ANOVA and the Student-Newman-Keuls (SNK) test. P-eNOS, total eNOS, and beta-actin protein levels were statistically analyzed in the same way. Fold changes in aspect ratio and angles of orientation were averaged and expressed as mean ± SEM. The statistical and mean comparisons tests used to analyze the aspect ratio data were ANOVA and SNK. To analyze alignment we used the t-test and Sidak-Bonferroni method to compare means within each range of angle orientation. F-actin fiber orientation within cell bodies was qualitatively analyzed.