Indeed, we have previously reported that culture of EC and fibroblasts inhibited the recruitment of PBL when they were in close contact on opposite sides of 3.0 μm pore filters, but not when 0.4 μm Stem Cell Compound Library mouse pore filters were used (McGettrick et al., 2010). To test how migration into 3-D matrix might be influenced by fibroblasts co-cultured with EC but not in direct contact, we modified the construct so that the EC were cultured on filters above collagen gels incorporating fibroblasts, with the two cell types separated by 600–800 μm (Fig. 1C). In this construct, we observed
similar adhesion of PBL to EC for mono- and co-cultures, with or without treatment with cytokines (data not shown). In the absence of cytokines, fibroblasts in the gel markedly increased the migration of PBL through the endothelial layer on the filter compared to EC Onalespib cultured alone, but this effect was much reduced when cultures had been treated with cytokines (Fig. 4A). Interestingly, however, fibroblasts significantly reduced the entry of the migrated PBL into the collagen gel, both in the untreated and cytokine-treated cultures (Fig. 4B). Of note, fibroblasts cultured within gels respond appropriately to cytokine-stimulation, up-regulating ICAM-1 expression and secretion of CXCL1 and CXCL10 to a similar level as that observed by fibroblasts cultured
on plastic (i.e. in the absence of collagen) (Supplemental Fig. 4). Moreover, these responses were maintained during co-culture with endothelium (Supplemental Fig. 4). Thus fibroblasts are capable of responding to cytokines and also suppressing T-cell entry into the gel, indicating a role for other factors in this effect. Thus, so far, fibroblasts tended to promote the migration of
PBL across EC when direct contact was prohibited, but tended to inhibit onward migration in co-culture. To gain insight into the latter effect, we examined the distribution of PBL and fibroblasts within the gels. The distances PBL migrated into the gels after 24 h were significantly reduced in the presence of fibroblasts for unstimulated or cytokine-treated cultures (Fig. 4C). Similar reduction in depth was also observed at 44 h (data AMP deaminase not shown). However, in examining the position of fibroblasts in the gel, we found that the depth of the gels was much less in their presence than in their absence (Fig. 4D). While we observed that fibroblasts were evenly distributed through the depth of the gel (data not shown), they had significantly contracted the gel. To evaluate the depth of penetration by PBL in a manner independent of gel depth, we calculated the proportions of PBL in the upper and lower halves of the gel. On average there were significantly more PBL in the upper half of the gel compared to the lower half (ratio about 60:40) (Fig. 4E). In addition, the proportion in the upper half was slightly higher (and the proportion in the lower half slightly lower) when fibroblasts were present in the gel.