Oriented matrix promotes directional tubulogenesis

Abstract

Detailed control over the structural organization of scaffolds and engineered tissue constructs is a critical need in the quest to engineer functional tissues using biomaterials. This work presents a new approach to spatially direct endothelial tubulogenesis. Micropatterned fibronectin substrates were used to control lung fibroblast adhesion and growth and the subsequent deposition of fibroblast-derived matrix during culture. The fibroblast-derived matrix produced on the micropatterned substrates was tightly oriented by these patterns, with an average variation of only 8.5°. Further, regions of this oriented extracellular matrix provided directional control of developing endothelial tubes to within 10° of the original micropatterned substrate design. Endothelial cells seeded directly onto the micropatterned substrate did not form tubes. A metric for matrix anisotropy showed a relationship between the fibroblast-derived matrix and the endothelial tubes that were subsequently developed on the same micropatterns with a resulting aspect ratio over 1.5 for endothelial tubulogenesis. Micropatterns in "L" and "Y" shapes were used to direct endothelial tubes to turn and branch with the same level of precision. These data demonstrate that anisotropic fibroblast-derived matrices instruct the alignment and shape of endothelial tube networks, thereby introducing an approach that could be adapted for future design of microvascular implants featuring organ-specific natural matrix that patterns microvascular growth.

Keywords: Anisotropy; Endothelial morphogenesis; Extracellular matrix; Micropattern substrate.

Figure 1. Patterned Fibronectin and WI38 Response

A micropatterned coverslip (with no cells seeded) was labeled for epifluorescence imaging using an anti-fibronectin antibody (a, c, e). The images shown correspond to the three regions examined in this study: vertical parallel lines (a), horizontal parallel lines (c), and the nonpatterned region (e). The red arrows in a and c denote the regions containing fibronectin. WI38 cells were seeded onto patterned substrates in serum free conditions and phase contrast images were collected at 2.5 hours (b, d, f). Representative phase contrast images of the cell attachment to the vertical region (b), horizontal region (d), and outside region (f) are displayed. The red arrows in b and d highlight regions where cells are attached and conformed to the parallel line patterns. The arrows in f identify cells adhered to the non-patterned region. Scale bar = 65 μm (panel a).

Figure 2. Fibroblasts on Patterned Substrate Over Time

Phase contrast images of WI38 cells on a horizontal pattern region (a, b, c) and a non-patterned region (d, e, f) after different times of culture are shown. Approximately the same location on the coverslip is shown for all images of the same sample. The etched number 30 is seen in a-c, and the same etched corner is seen in d-f. The left images (a, d) display the initial cell attachment to the coverslip after 2.5 hours, while the middle images (b, e) show a confluent monolayer of cells observed after 7 days of culture. The red arrows denote cells in these images. The fibroblasts were then extracted and the fibroblast-derived matrix was examined (right images, c, f). Inserts contain 1.5x magnification of matrix fibrils in the region marked with an *. The cells and their extracellular matrix maintained the horizontal alignment provided by the micropatterned lines (b, c). The same was true for vertical lines (data not shown). In the non-pattern region, the cells and their matrix aligned to some extent (e, f). However, the alignment is not uniform over the entire area. Scale bar = 65 μm.

Figure 3. Fibroblast-Derived Matrix in Linear Patterned Regions

Representative ECM produced by fibroblasts from vertical patterned (a), horizontal patterned (b), and non-patterned (c) regions are shown. Scale bar = 65 μm (panel a). Vector plots of the anisotropy metric and angles were represented in d-f. The average anisotropy of the fibroblast-derived matrix for the three regions examined here is shown in g. Both the vertical and horizontal pattern regions had significantly larger anisotropies than the outside region. ** indicates p<0.001. The preferred orientations are shown in h. Each dot represents a data point and the lines represent the mean angle for each region. The box and whiskers graph in i represents the angle of deviation in preferred orientation of the fibroblast-derived matrix fibrils from the original micropattern.

Figure 4. Endothelial Tube Formation in Regions of Linear Patterned Matrix

HUVEC formed tubes (indicated by blue arrows) when seeded on the fibroblast-derived matrix. Samples were labeled for F-actin (red), fibronectin (green), and nuclei (grey) (a merge of all three is shown in panels a-c). Representative data from vertical (a), horizontal (b), and non-patterned (c) regions are shown. Scale = 65 μm (located in a). Representative vector plots of the anisotropy and orientation of the HUVEC actin in the different regions are shown in d-f. The average aspect ratios of tubes from the three regions (vertical, horizontal and outside) are shown in g. Both the vertical and horizontal pattern regions had significantly larger anisotropies than the outside region. ** indicates p<0.001. The angles of the preferred tube orientations are shown in h. Each dot represents a data point and the lines represent the mean angle for each region. The box and whiskers graph in i represents the angle of deviation in preferred orientation of the endothelial cells from the original micropattern.

Figure 5. Variations in Micropattern Geometry

Coverslips containing micropatterns of “Y” shapes were studied (a and b). WI38 cells adhered to the pattern in serum free conditions at 2.5 hours and after 7 days produced matrix in accordance with the pattern (a, left image and b top left image) The WI38 adhesion and matrix production on a “Y” micropattern can be seen right above the number 78 (denoted with a blue arrow in each image). The corresponding amine labeled fibroblast-derived matrix image for that region is shown on the right. The insert in a is a 1.5x magnification of the region highlighted with the arrow. When HUVEC were seeded on matrices produced on these micropatterns, they formed tubes with a similar shape (b). The “Y” shape was documented at 2.5 hours at location 132 (denoted with a blue arrow). A branching tube was observed with phase at the same location above the 3 in grid 132 (blue arrow pointing at vessel). The corresponding epifluorescence images of fibronectin, actin, and an overlay with nuclei for this location are shown (blue arrows are a reference that marks the same location in all images in b). The inserts in b are a 2x magnification of the region highlighted with the arrow. Similar control of tube geometry was noted with fibroblast-derived matrix produced on micropatterns of fibronectin in the shape of “L” (c). The turn in the “L” micropattern was observed in grid 178 at 2.5 hrs. Endothelial tubes formed in a similar shape as at the same location (178). The corresponding epiflourescence images of fibronectin, f-actin, and composite with nuclei for this region are shown in. The blue arrows are a reference that marks the same location in all images in c. The inserts in c are a 1.5x magnification. The scale bars are 65 μm and the grid line width is 20 μm.