Our past studies on the mechanism of cyclic AMP (cAMP)-mediated control of tumor growth, using the experimental rat mammary tumor models as well as human breast cancer cell lines, indicated that the action of cAMP is mediated by the RII cAMP receptor protein, the regulatory subunit of cAMP-dependent protein kinase type II (Y. S. Cho-Chung, J. Cyclic Nucleotide Res., 6: 163, 1980). We now shown that the site-selective cAMP analogues, which are manyfold more active in binding to the cAMP receptor protein than previously studied analogues, demonstrate a potent growth inhibition of seven breast and three colon human cancer cell lines. The cAMP receptor protein has two different cAMP binding sites, and cAMP analogues that selectively bind to either one of the two binding sites are known as either site 1 selective (C-8 analogues) or site 2 selective (C-6 analogues). Nineteen site-selective analogues, C-6 and C-8 monosubstituted and C-6,-8 disubstituted, were tested for their growth regulatory effect. The majority of these analogues demonstrated an appreciable growth inhibition, with no sign of toxicity in all 10 cancer lines at micromolar concentrations. The three most potent inhibitors were 8-Cl-, N6-benzyl-, and N6-phenyl-8-thio-p-chlorophenyl-cAMP, demonstrating 50% growth inhibition at 5-25 microM concentrations (IC50). Furthermore, N6-analogues, in combination with halogen or thio derivatives of C-8 analogues, demonstrated synergistic enhancement of growth inhibition. The growth inhibition paralleled a change in cell morphology, an augmentation of the RII cAMP receptor protein, and a reduction in p21 ras protein. The growth inhibition by 8-Cl-cAMP was not due to its metabolite, 8-Cl-adenosine, since: (a) the growth inhibition by 8-Cl-cAMP was released upon cessation of treatment, whereas that by 8-Cl-adenosine was not released; (b) 8-Cl-cAMP treatment did not affect cell cycle progression, whereas 8-Cl-adenosine brought about G1 synchronization; (c) 8-Cl-cAMP treatment caused reduction of p21 ras protein, whereas 8-Cl-adenosine did not affect p21 levels; and (d) 8-Cl-adenosine was not detected in either cell extracts or medium from the cells treated with 8-Cl-cAMP for 48-72 h. Site-selective cAMP analogues thus provide a new physiological means to control the growth of breast and colon human cancer cells.