, 1988). In vivo, EC grow on a basement membrane in close juxtaposition with pericytes or smooth muscle cells depending on the vessel, but may not usually have direct contact with fibroblasts. Here, behaviour (morphology, recruitment of leukocytes and response to cytokines) of EC was not impaired when cultured on collagen matrix alone or as part of the double gel model (where EC are seeded above a gel, above a fibroblast containing gel). Such behaviour is similar to that observed when endothelial cells are cultured on a range of surfaces, including plastic tissue culture wells and Transwell filters (McGettrick et al., 2009a). This indicates that collagen itself is unlikely to impair EC function
or behaviour, rather the loss of integrity was a fibroblast-specific effect. The integrity of the endothelium may be differentially modulated by different stromal cells in vitro, and so the best model for co-culture might be different check details also. These findings do, however,
raise the question as to whether fibroblasts potentiated lymphocyte transmigration at the level of the filter rather than the endothelium in that model. This was investigated further in the layered-gel model (see below). In either model, fibroblasts reduced the proportion of transmigrated PBL that penetrated into the gel and those that did enter migrated only half as deep when fibroblasts were present. Of note, responses to cytokine-treatment were similar for fibroblasts cultured on plastic as those within the gel. In fact, higher levels of the adhesion molecules, ICAM-1, were observed in co-culture gel constructs, AZD2281 cell line indicating that the fibroblasts had sufficient
receptors to support and encourage lymphocyte migration through the gel. Density, spatial arrangement and source of collagen fibres have all been suggested to alter the ability of leukocytes Interleukin-3 receptor to move within gel constructs (Wolf et al., 2009). Here it became evident that fibroblasts caused significant contraction and reduction in depth of the gels. When we purposely made gels with increasing collagen concentrations, the inhibition of initial penetration was reproduced. Thus the main effect of the fibroblasts in the later stages of migration appeared to be through matrix modification, while effects through direct contact with the PBL or release of attractants were not obvious. The fibroblasts probably also deposited matrix proteins such as fibronectin over the duration of the culture and assay, and it would be interesting to investigate whether this might affect migration in the future. Preliminary studies where we have purposely added fibronectin into the collagen gels did not, however, cause increased penetration at least (G. Jevons; unpublished observations). Others have reduced fibroblast contraction of collagen gels through the chelation of divalent cations (e.g. Ca2 +) (Ilagan et al., 2010) or the antagonism of endogenous TGFβ signalling or heparin sulfate-containing proteoglycan synthesis (Chen et al., 2005).