We have reported previously that unsymmetrical disulfide inhibitors of the human thioredoxin/thioredoxin reductase redox system (hTrx/TR) possess antitumor activity. We have broadened the search for more potent inhibitors and evaluated a large range of mono- and bis-disulfide compounds, prepared using parallel syntheses. Reaction of isothioisourea-HCI salts (R') or bis-salts (R) with aromatic or aryl thiols (R") in wells of 96-well plates produced >450 derivatives with the structures R"SSR' and R"SSRSSR". The excellent yield and purity of the disulfides provided sufficient material for evaluations of enzyme inhibition and cytotoxicity. Selection criteria based on the IC50 values for hTrx/TR inhibition and for cytotoxicities of the disulfides identified agents for subsequent scale-up syntheses and in vivo evaluations of antitumor activity. These scale-up studies confirmed the original activities of agents synthesized in the plates and validated the parallel synthetic approach. Structure-activity information derived from the hTrx/TR IC50 data allow for a number of generalizations. The most potent inhibitors of the Trx system contained two heteroatoms ortho to the disulfide moiety in an aromatic functionality. The thioalkylating moieties had greatest activity with one branch point alpha to the disulfide. In the absence of branching, more potent inhibition was observed with the electron withdrawing functionalities. Bis-disulfides showed patterns of activity which depended on chain length, with optimum activity observed when the disulfide units were separated by 3.9 A, a similar distance to that separating the thioredoxin active site cysteine residues. From the agents selected for scale-up syntheses, three disulfide compounds were studied for their antitumor activity in vivo against human tumor xenografts in scid mice. One of the analogues discovered through the combinatorial syntheses/screening for Trx inhibition, 1-phenylethyl 2-imidazolyl disulfide, N1 (ProlX agent PX-C5), has demonstrated excellent in vivo activity against the MCF-7 human breast cancer and the HL-60 human leukemia, thus validating this approach for novel drug discovery.