Ideal restoration material for caries would allow attachment of gingival epithelia. The attachment of epithelial cells to specimens of the 4 most commercially used well- or partially-cured resin composites, with and without TEGDMA, was assessed. Effects of resin composite on the Ca9-22 gingival epithelial cell-line were assessed by measuring the cytotoxicity, viability and gene expression for attachment, apoptosis, ROS-production, pro-inflammatory cytokines, and matrix metalloproteinases. As controls, cells on tissue culture plastic or bovine tooth enamel specimens were used. Significantly less cell attachment was measured on freshly made resin-composite specimens. Concomitantly, significantly higher cytotoxicity was measured in the presence of freshly made resin-composite specimens. However, after 8 d of leakage, the cell attachment to and cytotoxicity of the resin composite was comparable to bovine tooth enamel. Significantly higher expressions of IL6, MMP2, BCL6 and ITGA4 were measured in cells attached to resin-composite surfaces than controls. There were no significant differences between the results using different conditions of resin composite, with or without TEGDMA and well or partially cured. Less cell attachment and presence of more inflammatory markers were observed on all freshly-made resin-composite surfaces. However, after a leakage period attachment of cells to the resin composite improved to the level of natural tooth materials such as enamel. This indicated that the negative effects of resin composites on epithelial cells might be transient.