Vascular Smooth Muscle Cell Durotaxis Depends on Substrate Stiffness Gradient Strength

Brett C. Isenberg, Paul A. DiMilla, Matthew Walker, Sooyoung Kim, and Joyce Y. Wong. Vascular Smooth Muscle Cell Durotaxis Depends on Substrate Stiffness Gradient Strength. Biophysical Journal, Volume 97, Issue 5, 2 September 1009, Pages 1313-1322

Doi: 10.1016/j.bpj.2009.06.021

Introduction:

Cell migration is important in many different biological processes including wound healing, angioneogenesis, cancer invasion, and growth and development. As to how cells migrate and what drives migration there are many different factors that have an effect, both inside and outside the cell. External signals can be both chemical, such as growth factor proteins, or physical, such as migration in the direction of flow above the cells. Cells can also be responsive to relative amounts of signal, such as in a gradient, rather than purely binary responses. The best understood gradient responses are for chemical gradients. For example, chemotaxis is the tendency for cells to respond to a gradient of soluble chemicals and haptotaxis is a gradient of adhesive ligands on the surface interface of the cell. However, purely chemical gradients aren’t the only gradients cells can respond to, as evidence of response to surface stiffness gradients, or durotaxis, have been seen in normally motile cells such as fibroblasts. Yet, durotaxis is not quantitatively understood in comparison to chemotaxis or haptotaxis, nor is the mechanism itself well characterized. This leaves many questions unanswered about durotaxis and much progress to be made.

Summary:

To attempt to quantify durotaxis and better understand the mechanisms involved, vascular endothelial cells were plated and tracked on gels of varying stiffness. This included gels of uniform stiffness and gels of gradient stiffness.

Gels were created using polyacylimide with a bis-acrylimide cross-linker to tune stiffness. The gradient gels were manufactured using a microfluidic device that appropriately mixed the cross linker in gradient fashion to create a stiffness gradient throughout the gel. The polymerization was UV light dependent. The uniform stiffness gels used ranged in stiffness from 1-100kPa and the gradients ranged from 1-4 kpa/ 100um. The absolute stiffness range was based on relevant tissue stiffnesses seen in human vasculature. The gradient range was based on vascular endothelial cells on average being 100um across.

Mechanical testing of the gels was done through use of an atomic force microscope (AFM). Indentations were made in triple on the uniform gels and in triple but also up and down the gradient of the gradient gels in 400 um intervals. The uniform gels showed a linear trend between cross-linker concentration and elastic modulus. The gradient gels did not show regular elastic behavior and in fact the AFM data could not be used, but rather a relative stiffness was calculated based on cross-linker slope of the linear portion of the curve from the uniform stiffness gels. (See figure 1)

Figure 1. Macro- and micromechanical properties of PAAm hydrogels fabricated with uniform bis concentration. (A) Bulk tensile modulus as a function of bis concentration from traditional uniaxial tensile testing. (B) Effective force constant, defined as the slope of the linear region of the indentation curve, as a function of bis concentration from AFM

Gels were functionalized with ECM ligand through a photo activated amine cross-linker to the poly-acrylimide. The only ECM protein used was collagen I. To ensure that haptotaxis was not a factor, fluorescent collagen I was adhered to the gels and scans ensured that independent of surface stiffness the ligand density was the same.

Cells used with bovine aortic vascular smooth muscle cells. Morphology was an important measurement taken and was collected and analyzed using Metamorph software along with ImageJ.

Cell tracking was done over 20 hours of incubation with selected cells being imaged every 15 minutes. The cell centroid location was measured and then all this data was collected an analyzed to determine the cells motility, including persistence.

For results, cell morphology was seen to vary with stiffness as cell spread further with higher stiffness. On the gradient gels more cells were spread than rounded compared to the uniform stiffness gels. Cell orientation was seen to align with the gradient on the gradient gels. Orientation of cells on uniform gels was seen to be independent of elastic modulus. (See figure 4 and 5)

Figure 4. Polarization and orientation for cells on uniform (A, C, and E) and gradient (B, D, and F) gels. (A and B) Percentage of cells with recognizable lamellipodia. (C and D) Average cell orientation with respect to an arbitrary reference direction for uniform gels (C) and gradient direction for gradient gels (D). A cell orientation of 0° indicates perfect alignment in the direction of the gradient; an orientation of 180° indicates perfect alignment in the direction opposite the gradient. Data labeled with ^* correspond to p <>E and F) Histograms of orientation angle. Data for 0 kPa/100 μm were based on pooling data for all uniform gels. Histograms were not statistically distinguishable. The † symbol indicates that only 10 cells were available for this condition, in contrast to >30 cells for all other condition

Figure 5. Scatter plots of cell orientation angle on gradient gels as a function of tensile modulus for different gradient strengths: (A) 1 kPa/100 μm, (B) 2 kPa/100 μm, and (C) 4 kPa/100μm. Vertical dotted lines delimit the range of moduli for individual gradients

Cell motility data showed that motiliy increased with stiffness to a point, at the higher stiffnesses cells would slow down. Also cells on the gradient gels were much more likely to follow the gradient and thus had higher persistence time. Additionally, cells on the steeper gradients showed greater preference to follow the gradient.

Figure 6. Windrose displays of typical paths of VSMCs over 20-h periods on uniform gels (A, top row) and gradient gels (B, bottom row). Arrows indicate direction of gradient from softer to stiffer region.

Critique:

There are a few issues I think stand out with this paper that take away from the significance of their data.

First, the motility data is not from many cells, less than 10 in some cases. This makes me question the statistical significance of this data.

Second, the stiffness of the gradient gels could not be verified directly and assumes that the microfluidic device used worked just as intended since they also could not directly measure the cross-linker concentration at any point. This needs to be characterized in better fashion. Also previous works have shown polyacrylimide gradient gels to in fact behave in elastic fashion and be easily measured using AFM. I would like to know more about why their gels aren’t elastic.

Finally, the use of just one ECM ligand seems like a limited approach. The actual cellular environment contains much more than just collagen and thus a collagen only gel or only looking at collagen gels is far removed from the in vivo environment. In particular previous studies have shown differential cell motility and persistence on different ligands, such as fibronectin.