Supplemental Data

Immunofluorescence Staining Methods

To obtain fluorescence images of F-actin and vinculin, ECs were fixed with 4% paraformaldehyde (Fisher Scientific) for 20 min and permeabilized with 0.2% Triton X-100 (Acros) for 2 min. ECs were then incubated in 0.5 μg ml-1 rhodamine-phalloidin (Sigma, St. Louis, MO) for 30 min. Non-specific protein binding was blocked by incubation with 1% bovine serum albumin (BSA, Sigma) for 45 min. After rinsing in 0.1% BSA twice, ECs were incubated in a 1:500 dilution of mouse monoclonal anti-vinculin (clone HVIN-1, Sigma-Aldrich) for at least 4 h at room temperature. After blocking in 2% normal goat serum (Jackson ImmunoResearch, West Grove, PA) for 30 min, ECs were incubated in a 1:100 dilution of Cy2-conjugated goat anti-mouse IgG secondary antibody (Jackson ImmunoResearch) for 1 h at room temperature. Coverslips were mounted on glass slides using SlowFade Gold anti-fade reagent (Molecular Probes, Eugene, OR).

VE-Cadherin staining was performed in a similar way. ECs were fixed and permeabilized as described above. After rinsing with 0.1% BSA twice, non-specific protein binding was blocked by incubation with 1% BSA for 30 min. Next, ECs were rinsed twice with 0.1% BSA and incubated in a 1:100 dilution of goat anti-human VE-cadherin (CD-144, clone C-19, Santa Cruz Biotechnology) for 2 h. ECs were washed in PBS and blocked in 5% normal mouse serum (Jackson ImmunoResearch) for 45 minutes. After rinsing with 0.5% mouse serum, ECs were incubated in a 1:100 dilution of CY3-congugated mouse anti-goat IgG secondary antibody (Jackson ImmunoResearch). Finally, the cells were rinsed, and coverslips were mounted using SlowFade Gold anti-fade reagent (Molecular Probes, Eugene, OR).

Micropatterned Substrates Control Cell Morphology and Structure

Bovine aortic ECs within a confluent monolayer exhibited a typical cobblestone morphology and polygonal cell shape (Fig. S1). In contrast, cells on 20-m wide micropatterned lines were more elongated in shape and aligned parallel to the direction of the lines. In order to measure cell shape elongation, shape index was computed as . Shape index values range between 0 for a line and 1 for a circle. The average shape index for ECs on micropatterned lines (0.33±0.11, mean ± S.D., n=109 cells) was significantly smaller (1-way ANOVA, p<0.05) than for ECs in an unpatterned confluent monolayer (0.68±0.11, n=110 cells), reflecting the elongated cell shape on micropatterned lines. The ratio of cell alignment was defined as the number of cells with oriental angle within 20 deg of the x-axis divided by the total number of cells. The alignment ratio increased from 0.12 for the unpatterned cells to only 0.97 for the cells on the micropatterns (Fig. S1). Thus, the geometric pattern on the substrate caused significant cell shape elongation and alignment.

ECs in a confluent monolayer on an unpatterned surface contained F-actin stress fibers that appeared randomly distributed basally and in peripheral bands (Fig. S1). In contrast, cells patterned on 20-m wide lines contained thickened stress fiber bundles at cell edges and cytoplasmic stress fibers that were oriented primarily parallel to the major axis of cell elongation. The ratio of aligned stress fibers was significantly greater for the patterned cells than for the unpatterned ones. These measurements show that micropatterned substrates cause the cells not only to take on an elongated morphology but also an aligned stress fiber orientation parallel to the micropatterned line direction.

In order to examine the distribution and morphology of focal adhesion sites at the termini of stress fibers interacting with the substrate, ECs were fixed and stained with a vinculin monoclonal antibody (Fig. S1). ECs in an unpatterned confluent monolayer contained focal adhesion sites that were randomly distributed throughout the cell and elongated focal adhesion sites oriented perpendicular to the cell periphery. In contrast, most focal adhesion sites in micropatterned ECs were aligned parallel to the major axis of cell elongation. There was a significant increase in the ratio of aligned focal adhesions in the micropatterned cells, compared with the unpatterned ones.

The cell density on the micropatterns was similar to that of the confluent monolayers, suggesting that the micropatterned cells were as closely packed as unpatterned confluent cells. Unlike single or subconfluent cells, ECs on the micropatterns did not have open spaces to which to migrate, and they were able to establish cell-cell contact with neighboring cells on the same pattern. VE-cadherin in the micropatterned cells was located at the cell boundaries in between cells, similar to that in confluent cells (Fig. S2).These attributes demonstrated that micropatterned cells were in a “quasi-confluent” state and therefore provided a basis for comparison between them and unpatterned confluent cells.

ECs on 115-m wide micropatterned lines exhibited morphology that varied with the distance from the line edges (Fig. S3). Cells near the centerline of the pattern had polygonal shapes similar to those in unpatterned confluent monolayers, whereas cells along the pattern edge were significantly elongated.

Figure Legends

Figure S1(A-F) Cell morphology (DIC), F-actin stress fibers, and vinculin-labeled focal adhesions in an unpatterned EC layer and in ECs on 20-μm lines of fibronectin. Scale bar, 20 m. (G-I) In micropatterned cells, the ratio of aligned to total cells, stress fibers, and focal adhesions were significantly increased on micropatterned lines that in confluent monolayers (2-9 fields of view; *p<0.05, t-test). Error bars: standard deviation.

Figure S2VE-cadherin (red) distribution in (A) unpatterned confluent monolayers of ECs and (B) quasi-confluent EC layers on micropatterned lines. Blue, nuclei. Scale bar, 10 m.

Figure S3Phase contrast image of quasi-confluent layers of ECs on micropatterned lines of fibronectin with width 115 m. Scale bar, 100 m.

Movie Legends

Movie S1Time-lapse of ECs in an unpatterned confluent monolayer during a 4-h period with no flow followed by a 16-h period after flow onset. Centroids of 4 example cells are marked (black dots before flow onset, white dots after flow onset) to illustrate triphasic mechanotaxis and locally correlated migration behavior. Frame-to-frame interval, 20 min. Flow direction, left to right.

Movie S2Time-lapse of ECs on a horizontal 20-m wide micropatterned line during a 4-h period with no flow followed by a 16-h period after flow onset. Centroids of 4 example cells to illustrate suppression of mechanotaxis after flow onset. The 4 example cells were migrating to the right before flow onset (black dots) in locally correlated fashion. After flow onset (white dots), locally correlated migration was disrupted and a preferred direction of migration was not established in 16 h. Frame-to-frame interval, 20 min. Flow direction, left to right.