Does negative selection involve accumulation of self-reactive thymocytes in thymic rosettes?

Thymic rosettes, the natural associations between thymocytes and either macrophages or dendritic cells, were isolated from the thymus by collagenase digestion and unit-gravity elutriation. Rosettes from mouse strains where either the V beta 6-bearing thymocytes are deleted because of reactivity with products of the Mlsa allele of the minor lymphocyte stimulating locus, or where V beta 17a-bearing thymocytes are deleted because of reactivity with IE class II MHC molecules, were compared with rosettes from appropriate control strains to test if a selective association with stromal cells preceded deletion. Rosettes from an Mlsa-bearing strain were able to stimulate an Mlsa-reactive T-hybridoma, but much of this stimulatory activity was attributable to the few B cells associated with the rosette preparations; the stromal components of the rosettes appeared to be poor presenters of Mlsa gene products. There was no enrichment of thymocytes bearing high or low levels of V beta 6 TcR in the rosettes from the Mlsa-bearing strain, which might have reflected the poor presentation by the stromal cells. However, nor was there detectable selective association of thymocytes bearing C beta 17a in the rosettes from an IE-positive mouse strain. This argues against binding and immobilisation on stromal cells as part of the deletion process, but not against the stromal cells delivering a rapid signal during a transient association, leading later to deletion.