In this work human platelet aggregation induced in vitro by ADP, collagen, arachidonic acid and U-46619 (a thromboxane A(2) analogue) was used as a functional test to characterize 19 anti-GPIIb (M series) and anti2 GPIIIa (P series) monoclonal antibodies whose epitope location is known for most of them. Additionally, flow cytofluorimetry was applied to study the epitope expression of these antibodies in resting, EDTA-treated and SFLLRN peptide (thrombin receptor agonist)-activated platelets. Antibodies M6 (epitope located at GPIIbH 657-665), P23-7 (GPIIIa 114-122) and P40 (GPIIIa 262-303) bind weakly to only 43%, 70% and 66%, respectively, of the resting platelet population. This binding was enhanced in EDTA-treated and in activated platelets. Platelet activation enhances the apparent binding of most of the other antibodies. Further evidence on the existence of agonist-specific activated states of GPIIb/IIIa was provided by the agonist-dependent immunochemical inhibition in vitro of platelet aggregation by some of the anti-subunit antibodies studied here. The most notable cases are those of P40 and M6, which at 140 nM inhibit most, the platelet aggregation induced by arachidonic acid and U-46619. On the other hand, three of the most strong and agonist-independent inhibitors, P37 (GPIIIa 101-109), P97 and P95-2 (GPIIIa N-terminal half) bind to resting platelets with high affinity (5-8 nM), compete with each other for binding to GPIIb-IIIa and their epitopes are located at the N-terminal domain of GPIIIa, where the receptor ligand binding site(s) have been found. Given that the formation of activated GPIIb-IIIa (GPIIb-IIIa*) is the first step at which the anti-subunit antibodies can intervene as inhibitors and that agonist-specific inhibitors should block only agonist-specific steps, while nonspecific inhibitors should block steps common to all the agonists, then our present work support the hypothesis that there are different agonist-specific GPIIb-IIIa*s or, alternatively, different receptor environments, that can be specifically blocked by some of the antibodies. These results add to earlier evidence on agonist-dependent ligand specificity and activated states found for this and other integrins. Finally, the correlation between the in vitro inhibition of platelet aggregation and the antithrombotic activity in vivo is discussed for these antibodies.