The influence of crosslinking agents and diamines on the pore size, morphology and the biological stability of collagen sponges and their effect on cell penetration through the sponge matrix.

Artificial skin substitutes based on autologous keratinocytes cultured on collagen substrata are being developed for treating patients with severe burns. The properties of the collagen substrate can be manipulated, for example, by crosslinking, to optimize desirable properties such as cell growth and penetration into the substrate, biological stability and mechanical strength. Collagen sponges crosslinked with 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide (EDAC) and the diamine, diaminohexane, were used to determine the effect of crosslinking on pore size and morphology, on the stability of the crosslinked sponges both in cell culture media and during incubation with collagenase, and on the penetration of keratinocytes and fibroblasts through the sponge matrix. Crosslinking of the sponges reduced the pore size, particularly at the surface, and altered sponge morphology. After crosslinking the collagen fibers were thinner, and appeared lacy and delicate. Crosslinking also influenced sponge stability. In keratinocyte serum-free medium the pore size of plain collagen sponges increased with increasing incubation time, and crosslinking appeared to prevent this, and may have stabilized sponge structure. Incubation in serum-containing Dulbecco's minimum essential medium caused a marked reduction in pore size in both plain collagen and crosslinked collagen sponges. Crosslinking did not appear to influence this cell-free contraction of collagen sponges. Treatment of sponges with EDAC markedly increased the resistance of sponges to collagenase digestion. The penetration of both keratinocytes and fibroblasts was retarded by crosslinking the sponges. Fibroblasts penetrated through the sponges to a greater extent than keratinocytes, and their proliferation rate was faster. The total number of cells populating the crosslinked sponges after 10 days culture was approximately 50% of that on untreated collagen sponges. The mechanism responsible for this effect was different with the two crosslinkers used. Diaminohexane appeared to inhibit cell growth, whereas EDAC may have caused a decrease in cell adhesion to the sponges, without an apparent inhibition of growth rate. In terms of morphology, fibroblasts were elongated to a greater extent on crosslinked sponges, and alligned themselves along the collagen fibers. Keratinocytes grew in colonies on untreated sponges, but on crosslinked sponges they grew in isolation, with minimal cell-cell interactions. It may be necessary to reach a compromise to obtain the best combination of properties for using collagen sponges as substrata for artificial skin substitutes. Copyright 2001 Kluwer Academic Publishers