The potential of activated HLA class II-positive T cells as antigen-/alloantigen-presenting cells remains controversial. In our model system we use in vitro-primed, HLA class II-specific T cells of the memory T-cell phenotype, CD4+, CD29+ (4B4+), and CD45RO+ (UCHL-1). We have previously shown that alloactivated, HLA class II-positive T cells (Ta) are unable to stimulate proliferative responses in naive and primed allospecific T cells when 'back-stimulation' is avoided. The explanation of this feature of Ta is unknown, but it is due neither to suppression nor to insufficient HLA class II expression. Accordingly, we investigated the possibility that Ta have a deficient expression of accessory signals critical for the induction of proliferative T-cell responses. We found that (1) non-mitogenic concentrations of phorbol myristate acetate (PMA) in combination with either rIL-4, a CD28-reactive MoAb (Kolt-2), or a calcium ionophore (A23187) enabled Ta to elicit alloantigen-specific memory T-cell responses and to present purified protein derivative (PPD) to PPD-specific T-cell lines. The addition of irradiated, Epstein-Barr virus-transformed B-cell lines (EBV-LCL) (but not their supernatants) had a similar but less pronounced effect; (2) MoAb directed against HLA class II, CD25 (IL-2R), CD2, CD4, CD11a (LFA-1), or CD45RO molecules inhibited these responses; (3) PMA was required within the first hour of culture in order to induce optimal alloantigen-specific T-cell activation, while rIL-4 was fully effective when added after 20-44 h of culture; (4) incubation for 20 h of Ta with rIL-4 plus PMA markedly up-regulated CD54 expression on the Ta, and IL-4 seemed to potentiate the effect of PMA on the CD54 expression. In conclusion, the present data indicate that the inability of Ta to elicit (allo)antigen-specific, proliferative T-cell response is due to a lack of critical accessory signals. Up-regulation of CD54 was not sufficient for Ta to stimulate proliferative responses. Neither cytokines (IL-1, IL-6, and others) nor triggering of CD2 epitopes (T11.2 and T11.3) by soluble MoAb or solid phase support by MoAb against a number of accessory molecules provided the necessary signals. Thus, our data indicate that other, as yet unknown, signalling pathways play a key role in antigen- and alloantigen-specific T-T interactions. These pathways still need to be identified.