I Write Too Much

Remember back in high school, when everybody tried to inflate the size of their papers by narrowing the margins, using fonts that were too big, and trying to add extra spacing in? And how in college, that mostly went away because paper sizes started to feel more natural? Well, So far, grad school has me widening margins, decreasing font sizes, and eliminating line spacing. I fit this onto two pages. (Go ahead and read it if you like that sort of thing.)





Challenges continue in the field of creating and sustaining ideal environments in which new cells can grow and adapt. Of particular interest is the study of how to grow or develop tissues, starting with existing tissue fronts (such as at a wound site) or some number of initial cells (such as seeded, undeveloped cells). This paper seeks to summarize some of the promising methods of tailoring specific environments to meet cell developmental needs currently being researched.

A promising branch of the ongoing work looks into moving past the study of specializing cells via traditional stand-alone chemical developmental factors. Additional control can be applied using properties of the extracellular matrix to create additional or complimentary effects. This is a promising field of study both because of the potential to base the matrix on a wide variety of materials, and also for its potential to tailor those materials to naturally contain one or more desired factors to influence cell development. (Hileshorn, 2007)

Liu, Heilshorn, and Tirrell (2004) studied endothelial cell adhesion on customized protein-based extracellular matricies containing RGD and CS5 cell-binding domains. Using protein-based scaffolds is just one of many options for studying cell/environment interactions. Poly(ethylene terephthalate) and expanded poly(tetrafluoroethylene) are cited in the current study as alternative substrates out of which to make implanted synthetic vascular grafts. By creating a protein matrix on which to mount the cells, the option of adding in the amino acid RGD and CS5 binding domains to test their relative adhesive potential for endothelial cells became possible. Since the RGS and CS5 domains are themselves amino acid sequences, they are but two of many building blocks available for the construction of the protein matrix scaffold.

The fact that these domains are naturally occurring in fibronectin and are thus familiar hints at an advantage that protein-based scaffolds boast: extreme customizability. The ability to create cellular factories from bacterial cells, all producing amino acid chains of identical length and pre-tailored features offers major advantages for study. Put directly, the ability to exactly replicate amino acid chains of the same length is a degree of control not seen in other matrices, such as the polymer substrates mentioned earlier. The ability to exactly duplicate amino acid sequences of specific functional groups in sequence is also a feature unique to the protein scaffold. The ability to study one or a few functional groups at a time allows for a great deal of control for research purposes. However, in the long term, protein-based scaffolds may offer the most tailorable extracellular environment simply due to the naturally-occurring wide variety of functional groups available in the protein chains. (Hileshorn, 2007)

That said, it is worth noting that there are drawbacks to protein-based scaffolds. The very mechanism of production – bacterial cell construction – has a strong potential to add contaminants into the isolated protein chains. To isolate the grown proteins, cells are typically lysed and the resulting mix then purified. Bacterial cells generally inspire an immune response in living creatures, thus, contaminants that would be naturally found as parts of the lysed cells may result in dangerous side effects during practical use of the desired protein chains, unless they are very well isolated. Additionally, the batch yields of bacterially-produced proteins are much smaller than those of many alternative scaffolds. (Hileshorn, 2007)

Returning to the topic of vascular grafts, the protein-based scaffold was chosen to test the feasibility of overcoming limitations in the artificial materials currently used. Liu, Heilshorn, and Tirrell (2004) report that the poly(ethylene terephthalate) and expanded poly(tetrafluoroethylene) grafts are prone to failure due to thrombosis and intimal hyperplasia. In other words, the grafts are often seen to either restrict down due to unnatural clotting on the inner wall of the graft, or the graft inspires cell growth to the point that living tissue restricts the flow of blood through the passageway itself. The exact mechanism via which these responses are generated is not perfectly understood, but use of the protein scaffold was chosen for study as an alternative to try to compensate for two possible factors. First, the rigidity of the protein matrix can be customizable to more closely mimic those seen in healthy arteries. Engler, et al. (2006) demonstrate that the elasticity or relative stiffness of the matrix on which cells bind has a pronounced effect on the eventual cell specialization and characteristics. Thus, creating a graft which more closely moves and bends like a healthy artery would be expected to develop cells more typical of those living in a healthy artery. Second, the protein substrate was tailored to specifically incorporate the RGD and CS5 binding domains, in the hope that they would encourage adherence of endothelial cells like those found lining the inner walls of healthy arteries. Having this endothelial cell lining would offer artificial vascular grafts the added benefit of duplicating the expected environment for permeation by the immune system and other processes normally carried out by arterial linings (as opposed to the hyperplasia otherwise often seen).

However, it must be noted that initially, there was no specific proof that the use of the RGD and CS5 binding domains were specifically responsible for favorable endothelial cell adhesion. Without further study, some unknown aspect of the protein matrix could have been arguably just as (if not more) responsible for any and all results. Thus, Liu, Heilshorn, and Tirrell (2004) tested multiple variables via which they could determine the actual results of the use of the binding domains themselves.

First, four artificial extracellular matrices were used to mount cells, and then were stress-tested by attempts to centrifugally detach the deposited cells after they were allowed to bind. Two of the four matrices were proteins containing RGD and CS5 domains, respectively. The second two matrices were identical, with the exception that the RGD and CS5 domains had two of their amino acids switched, and thus would no longer act as RGD and CS5 domains (though they would remain otherwise extremely similar). These studies showed that the matrices containing the proper binding domains contained higher observed cell adhesion indices than their incorrect counterparts as seen by a roughly 60% increase for the RGD domain and a roughly 100% increase for the CS5 domain adhesions.

Second, the cells adhering to the four substrates (the same as in the first test) were allowed to spread (settle onto and bind with) the surface of the four substrates over a period of time. In these tests, the cells on the correct RGD domain substrates spread similarly to those on the positive control fibronectin. (This was a qualitative trial, with cell shape being the observed variable.) However, the cells on all other substrates did not show fibronectin-like spreading – including those on the correct CS5 domains. From those results, it is difficult to conclude anything about the CS5 binding domain’s effect on endothelial cell adhesion. However, the RGD domains do seem to be either binding with the cells as expected or in some other way inducing the cell shape to allow for greater contact between the cell and the substrate (and thus have more area over which for other binding mechanisms to apply).

A third trial was performed on the two substrates featuring the proper RGD and CS5 domains. In this test, cells were allowed to settle onto the substrates, but then, peptides were introduced into the system which would themselves bind – in competition with the proteins in the cells themselves – with the binding domains in question. In both cases, the competing peptide had no pronounced effect on cells on control fibronectin substrates (proving that the peptides introduced do not naturally remove the cells via some innate property), but dramatically reduced or eliminated the number of adhering cells on the RGD- and CS5-only domains. In other words, the specific competition for these binding domains seemed to – as expected – eliminate the cause of cell adhesion to the substrates. There is one curious note to this set of tests, however. In the trails examining the effects of modified competing peptides (which would be expected not to be able to bind/compete), it was shown that the addition of modified (inactive) peptides similar to those competing for the CS5 binding domains, the added (inactive) peptides actually nearly doubled the number of observed adherent cells. This is an unexpected result, because if the peptides were truly having no specific interaction (which is precisely what they were selected to do), they would be expected to have a negligible effect on the system, much as is seen in the RGD trial with similarly inactive peptides. While this finding does not directly argue with the stated conclusion that the binding domains are the cause of cell adherence in these tests, it does suggest that some other force(s) may be at work as well.

Given that cell binding appeared to be taking place via the binding domains – especially on the more clearly confirmed trials for the RGD domains – a fourth trial was performed to examine the internal characteristics of the cells bound specifically to the RGD substrates. Here, cells were dyed to show F-actin and vinculin filaments – cell proteins found at focal adhesions, such as those which would be present at bound RGD domains. As expected, cells on the RGD substrate showed comparable F-actin and vinculin filaments forming when compared to those found in cells on fibronectin. These fibers were not seen in cells on negative control substrates, which, again, is strong evidence that specific, localized binding domains (such as the predicted RGD) are causing the cell adhesion.

It is further interesting to note that Engler et. al. (2006) have looked further at these cell focal adhesions and their effects on cell development. In their work, they discovered that a cell is capable of sensing the mechanical properties of the substrate on which it is growing (rigidity, tensile strength, etc.), and that these substrate properties have strong effects on the eventual specification of undeveloped cells as they adapt to more specialized tasks/environments. This raises the question of how exactly the cell is able to sense, let alone respond to the mechanical properties of the scaffold. In their research, they propose that the same actin filaments stained and examined in the RGD domain binding work (Liu, Heilshorn, and Tirrell, 2004), along with their focal adhesion points (including the RGD complexes) provide a structure capable of being pulled and exerting forces in response to the movements of the extracellular matrix (previously called the substrate). Engler et. al. point out that these same actin/focal adhesion complexes are known to interact with signaling molecules that may act as the transducers of this mechanical stimulus. This adds further interest to the question of artificial vascular grafts. Given the modified structural properties of the current artificial grafts, the cells may not only be failing to settle onto expected cell-binding domains, but may also be adapting to an entirely inappropriate (non-vascular) task in the body due to responses to the mechanical properties of the artificial matrix.






References



Engler, Adam J.; Sen, Shamik; Sweeney, H. Lee; Discher, Dennis E. Cell, Volume 126, August 25, 2006, pgs 677-689. “Matrix Elasticity Directs Stem Cell Lineage Specification.”


Heilshorn, Sarah. Lecture, March 28, 2007. “Engineering the Cellular Microenvironment: Protein-Based Scaffolds, Patterned Substrates, and Microfluidic Devices.”



Liu, Julie C.; Heilshorn, Sarah C.; Tirrell, David A. Biomacromolecules 2004, Volume 5, pgs. 497-504. “Comparative Cell Responses to Artifical Extracellular Matrix Proteins Containing the RGD and CS5 Cell Binding Domains.”