We sought to determine whether osteoblasts (OBs) can serve as accessory cells (ACs) for T-cell activation and whether T cells directly activate OB production of IL-6, using primary human OBs (NHOst), the transformed fetal osteoblast line hFOB1.19, and an osteosarcoma line SaOS-2. Robust, bidirectional activating interactions were shown using each of these three human ostoblast lines. BACKGROUND Osteoblasts (OBs) could come into contact with lymphocytes during inflammatory joint destruction and fracture repair. METHODS We used several in vitro assays to assess the ability of T cells and OBs to interact in the generation of immune and inflammatory responses. RESULTS By flow cytometry, three OB cell lines all were found to express ligands for T-cell co-stimulation. The integrin ligand CD54/ICAM-1 was constitutively expressed by hFOB1.19 and NHOst and was upregulated on SaOS-2 by IFN-gamma. MHC Class II was upregulated on all three lines by IFN-gamma. CD166/ALCAM, a ligand of the T-cell molecule CD6, was constitutively expressed on all three lines. A second putative CD6 ligand designated 3A11 was expressed on hFOB1.19 and NHOst, but not consistently on SaOS-2. The ectoenzyme CD26 (dipeptidyl peptidase IV) was expressed on hFOB1.19 and NHOst, but not on SaOS-2. All three cell lines presented superantigen to T cells, especially after treatment with IFN-gamma. Superantigen presentation was inhibited by antibodies to the leukocyte integrin CD11a/CD18 (LFA-1), MHC Class II, and CD54/ICAM-1. T cells, particularly when cytokine activated for 7 days before co-culture, stimulated all three osteoblast lines to produce interleukin (IL)-6, and this effect was boosted when IL-17 was added to the co-cultures with either resting T cells or cytokine-activated T cells. CONCLUSIONS Bidirectional activating interactions are readily shown between human T cells and several types of human OBs. The expression by OBs of ligands for the T cell-specific molecule CD6, as well as other molecules involved in immune interactions, strongly suggests that such in vitro interactions are representative of physiologic or pathologic events that occur in vivo.