Scratch–wound assay for EGF-induced migration of human dermal fibroblasts. Photomicrographs of representative regions of scratch–wounds at the indicated times. Vertical lines indicate the alignment of the 0 h image and the 49 h image. Scratch-wounded cultures of fibroblasts were incubated with medium that contained the indicated growth factor: photoprotected (caged) polypeptide, caged polypeptide following photolysis, recombinant (commercial) EGF, or no growth factor. In all experiments that contained growth factor, it was present at 1 ng mL À 1 , which was chosen because the 

Context in source publication

Context 1

... are difficult to erase or amend. Thus, there is a need for methods that allow dynamic, high-resolution creation of growth factor gradients without dependence on a particular substrate. To fill this need, we are exploring the use of caged growth factors. Photocaging is an approach to controlling protein activity with high precision in space and time. [20–32] In this approach, a photolabile protective group masks a critical functional ele- ment of the protein, yielding an inactive species. Photolysis of the masking group “uncages” the protein in its active composi- tion. This activation can be controlled with the high spatial precision with which light can be directed ( < 1 m m). [33] Thus, we consider creation and utilization of caged growth factors to be a promising approach to directing cell behavior. A photo- caged factor that is delivered to a two- or three-dimensional culture of cells can be activated by photolysis in a defined region at the desired time and washed away when its effects are no longer desired. Native human epidermal growth factor (EGF) is a polypeptide of 53 residues (Scheme 1 A). [34] It has mitotic, chemokinetic, and chemotactic effects on a wide range of cell types. We have created a light-activated polypeptide by synthesizing the EGF sequence with a photolabile protective group on the side chain of glutamate 40. Mutation of this side chain to a gluta- mine has been shown to result in a fourfold decrease in receptor-binding activity. [39] Thus, we speculated that placement of a protective group on this glutamate would interfere with receptor binding by masking the carboxylate side chain of that resi- due. We synthesized the protected peptide by incorporation of a photoprotected glutamate monomer (Scheme 1 B) during solid-phase peptide synthesis. The caged glutamate derivative that we employed was stable to the conditions of Fmoc solid- phase peptide synthesis. The presence of the a -nitropiperonyl protective group in the full-length product was confirmed by matrix-assisted laser desorption ionization mass spectrometry (MALDI MS). Removal of this protective group was observed chromato- graphically (Figure 1 A). The HPLC trace of the starting material was a double peak with a shoulder, despite multiple rounds of HPLC purification. The double peak might be due to the presence of two poorly resolved diastereoisomers originating from the racemic protective group. The persistence of the shoulder through multiple rounds of purification suggests that it is an alternatively folded form of the polypeptide in equilibrium with the predominant fold. Irradiation with near-UV light (365 nm) produced a single product peak that appeared at approximately the same rate as the disappearance of the starting material, with a half-life of 36 s at 30 mW cm À 2 (Figure 1). The molecular weight of the photolysis product was determined by MALDI MS and was consistent with removal of the protective group (see the Supporting Information). The following experiments demonstrate that the photolyzed material has biological activity that is qualitatively similar to that of EGF. To assess the effect of photocaging on the mitogenic activity of the polypeptide, we measured the factor’s ability to stimulate proliferation of fibroblasts. The caged factor was added to a culture of fibroblasts, and it was added after complete photolysis to a separate culture of fibroblasts. In control experiments, a commercially obtained sample of native recombinant human EGF or no growth factor was added to separate cultures. The mean cell density in each culture as a function of time was determined (Figure 2 A). The caged polypeptide did not stimulate proliferation of the fibroblasts at the concentration tested (50 ng mL À 1 ); the cell density decreased slightly as a function of time to an extent similar to that in the factor-free negative control. However, after photolysis, the uncaged syn- thetic growth factor stimulated proliferation of the fibroblasts to an extent similar to that stimulated by recombinant EGF. The concentration dependence of polypeptide-stimulated proliferation before and after photolytic uncaging is shown in Figure 2 B. A portion of a stock solution of protected EGF was photolyzed to completion. Photolyzed and unphotolyzed factor were then added at varying dilutions to cultures of fibroblasts, which were incubated for 72 h prior to determination of the mean cell density. At sufficiently high concentrations ( > 80 ng mL À 1 ), both the caged polypeptide and its photolysis product stimulated proliferation of fibroblasts. However, over the concentration range of 25–50 ng mL À 1 , the unphotolyzed growth factor displayed little or no activity, whereas the photolyzed factor displayed near-maximal activity. We have tested the activity of the caged polypeptide before and after photolysis for stimulating migration of human dermal fibroblasts. Cell migration was observed qualitatively with a “scratch–wound” assay [40, 41] (Figure 3). A confluent culture of fibroblasts was scratched to create a cell-free region. Cell movement was observed as the cells migrated to fill in the scratch. The caged polypeptide stimulated little migration, comparable to a control experiment with no growth factor. However, the same material after photolysis stimulated sub- stantially more “healing” of the scratch. Cell migration was assessed quantitatively by using the Boyden chamber method, [42] in which the number of cells migrating through a porous membrane is determined. Different initial growth factor concentrations on opposite sides of the membrane create a gradient to which the cells can respond chemotactically or chemokinetically. The chamber in which the cells were seeded initially contained growth-factor-free medium, and the chamber on the other side of the membrane initially contained 1 ng mL À 1 polypeptide. Parallel experiments were done with caged polypeptide, caged polypeptide after complete photolysis, and in control experiments, recombinant native EGF, or no growth factor (Figure 4). Consistent with the scratch–wound assay, the caged polypeptide is almost completely inactive in the Boyden chamber assay; the number of cells migrating in its presence are within experimental error of the number of cells migrating in the absence of any growth factor. Furthermore, photolysis of the caged polypeptide resulted in a substantial increase ( P < 0.0002) in the ability to stimulate cell migration. Though MS and HPLC analysis indicated that photolysis had been carried out to completion, the photolyzed material did not stimulate migration as effectively as the fivefold increase in cell migration brought about by native EGF. Nevertheless, photocaging clearly affords a method for triggering its ability to stimulate cell migration. Whereas many of the effects that have been sought previously with light-triggered molecules are rapid, the large-scale phenotypic effects of growth factors, are apparent after time periods of hours to days. The active factor must be present, at least intermittently, [43] during this period. Therefore, to spatially control the effects of the light-activated polypeptide, it was neces- sary to maintain a gradient in its concentration over a time period of that magnitude, despite diffusion on a more rapid timescale. Continuous activation of the growth factor in the desired region was not possible, because continuous irradia- tion with near-UV light at relevant power levels killed the cells (data not shown). Therefore, we repetitively established a gra- dient by intermittently activating growth factor in the desired region. Fibroblasts were seeded uniformly on a glass window of the culture device and allowed to adhere and spread. Defined medium that contained caged polypeptide (40 ng mL À 1 ) was delivered to the culture, and a region of the culture was illuminated intermittently. Illumination times for activation (see cap- tion of Figure 5) were chosen based on the relationship of con- centration to proliferative activity in Figure 2 B, the expectation that the photolysis rate would be linearly related to the light power, [44] and an estimate of the diffusion coefficient of the EGF in aqueous medium of 10 À 6 cm 2 s. [45] Preliminary experiments confirmed survival and normal proliferation of the cells with the chosen light regimen (data not shown). In our model for this process, as activated growth factor diffused out of the illuminated area, thereby diminishing the gradient, caged growth factor diffused into the illuminated area from the sur- rounding volume of the culture chamber and became available for photolysis. As the caged factor was depleted, the gradient that could be produced became shallower, so the medium was replaced regularly. The illumination and fluid replacement were automated to make the procedure reproducible and conven- ient. The result after three days is shown in Figure 5. A clear band of cells (ca. 300 m m wide) is visible in the position of illumination. The nonuniformity of the cell density within this band is comparable to that typically seen in nonconfluent cultures of this cell type. A general alignment of cells along the edge of this band is noteworthy, though the origin of this effect is not clear. Outside of the illuminated region, few spread cells remain, and a low density of spheroidal cells decreases with increasing distance from the illuminated region. In control experiments, no illumination-dependent variations in cell distri- bution were observed when recombinant EGF was used instead of the caged polypeptide. The cells remained spread and proliferated uniformly across the surface. Also, when growth factor was omitted entirely from the medium, few cells sur- vived for three days, instead a low density of spread and spheroidal cells was distributed without illumination-dependent variation. When the light-activated factor was used, but cells were omitted from the chamber until after the ...