Functional and molecular studies of in vivo-activated T lymphocytes involved in normal and abnormal immune responses, i.e., cells infiltrating tissues affected by autoimmune processes, require their previous in vitro expansion. Problems such as unavailability of specific antigen(s) (Ag) and/or the requirement of large amounts of autologous peripheral blood mononuclear cells (PBMC) as feeders have prompted the development of alternative expansion methods that circumvent the use of antigen-presenting cells (APC) and/or Ag. We have previously shown that cytomegalovirus (CMV)-specific T cell lines and clones can be efficiently propagated in long-term cultures by stimulation with agonistic monoclonal antibodies (mAb) coated onto polystyrene particles in the absence of APC. However, this and other stimulation protocols may skew the repertoire of clonotypes that proliferate in response to nominal Ag and APC. Here we show that polyclonal populations of CMV-primed and interleukin-2 (IL-2)-stimulated PBMC undergo the same clonotypic selection when induced to grow both by continuous stimulation with CMV and an anti-CD3 mAb presented by APC. This selection, reproduced in three independent expansion experiments, involved the dominant growth of two CMV-specific, IL-2-secreting CD4+ clonotypes sharing J alpha, J beta, V alpha and V beta genes and highly homologous T cell receptor (TcR) junctional sequences. The dominant growth of these 2 clonotypes required a direct T cell/APC interaction since when coated onto polystyrene particles the same mAb induced the selective expansion of another IL-2-secreting CD4+ CMV-specific clonotype expressing a different, yet homologous, TcR heterodimer (used same V alpha and V beta genes), which was underrepresented before expansion (5 vs. 58%). T cell clones belonging to the subdominant clonotype proliferated significantly faster in response to stimulation with anti-CD3 mAb coated onto beads than in response to stimulation with CMV or anti-CD3 mAb presented by APC, possibly due to long-term expansion without APC or antigen. In contrast, neither the dominant clonotypes nor unprimed T cells were able to undergo CD3-mediated expansion in the absence of APC. We conclude that quantitative differences in growth competency among normal Ag-activated T-helper (Th) clonotypes in response to antigenic stimulation can be reproduced by stimulation through the TcR in the absence of TcR occupancy but only in the presence of APC and that certain clonotypes do not require APC for long-term growth in vitro. These data have practical implications for the isolation and repertoire characterization of in vivo-activated T cells from tissues affected by inflammatory, i.e. autoimmune, phenomena.