Site-selective cyclic AMP analogs bind to site 1 or site 2 of the known cAMP-binding sites depending on the position of substituents on the purine ring, either at C-2 and C-8 (site 1) or at C-6 (site 2). The growth inhibitory effect of such site-selective cAMP analogs used in this investigation with 15 human cancer cell lines surpassed that of analogs previously tested. The most potent analogs were 8-chloro, N6-benzyl and N6-phenyl-8-p-chlorophenylthio-cAMP. The combination of a C-8 with an N6 analog had synergistic effects. The 24 site-selective analogs tested produced growth inhibition ranging from 30 to 80% at micromolar concentrations with no sign of toxic effects. Growth inhibition was not due to a block in a specific phase of the cell cycle but paralleled a change in cell morphology, an increase of the RII cAMP receptor protein and a decrease of p21 ras protein. Since the adenosine counterpart of the 8-chloro analog produced G1 synchronization without affecting the RII and p21 ras protein levels, it is unlikely that an adenosine metabolite is involved in the analog effect. Site-selective cAMP analogs thus provide a new biological tool for control of cancer growth.

The physiologic role of cyclic adenosine monophosphate (cAMP) in the growth control of a spectrum of human cancer lines, including leukemic lines, and v-rasH oncogene-transformed NIH/3T3 cells is demonstrated by the use of site-selective cAMP analogs. These cAMP analogs, which can select either of the two known cAMP binding sites of the cAMP receptor protein, induce potent growth inhibition, phenotypic change, and differentiation (leukemic cells) of cancer cells at micromolar concentrations with no sign of cytotoxicity. The growth inhibition parallels selective modulation of cAMP-dependent protein kinase isozymes, type I versus type II, and suppression of cellular proto-oncogene expression. Site-selective cAMP analogs thus provide new biological tools for investigating cell proliferation and differentiation and also for the improved management of human cancers.

Two classes (site 1-and site 2-selective) of cAMP analogs, which either alone or in combination demonstrate a preference for binding to type II rather than type I cAMP-dependent protein kinase isozyme, potently inhibit growth in a spectrum of human cancer cell lines in culture. Treatment of K-562 human leukemic cells for 3 days with 30 and 10 ,uM 8-chloroadenosine 3',5'-cyclic monophosphate (8-Cl-cAMP) (site 1-selective) resulted in 60% and 20% growth inhibition, respectively (with over 90% viability). N'-Benzyl-cAMP (site 2-selective) (30 IpM) treatment resulted in 20% growth inhibition by day 3. When 8-Cl-cAMP (10 pM) and uM) were both added, growth was almost completely arrested. The growth inhibition was accompanied by megakaryocytic differentiation in K-562 cells. The untreated control cells expressed little or no detectable levels of glycoprotein nwb-rn surface antigen complex. 8-Cl-cAMP (30 ,uM) treatment for 3 days substantially increased the antigen expression, while N6-benzyl-cAMP caused little or no change in the antigen expression. When cells were treated with 8-Cl-cAMP in combination with N6-benzyl-cAMP, antigen expression was synergistically enhanced, and cells demonstrated megakaryocyte morphology. By Northern blotting, we examined the mRNA levels of the type I and type I protein kinase regulatory subunits (RI. and Rllp), the catalytic subunit, and c-myc during 8-Cl-cAMP treatment. The steady-state level of RUp, cAMP receptor mRNA sharply increased within 1 hr of treatment and remained elevated for 3 days, while that of the RIU receptor markedly decreased to below control level within 6 hr and remained low during treatment. However, 8-Cl-cAMP did not affect the mRNA level of the catalytic subunit. 8-Cl-cAMP treatment also brought about a rapid decrease in c-myc mRNA. Thus, differential regulation of cAMP receptor genes is an early event in cAMP-induced differentiation and growth control of K-562 leukemia cells.A major defect ofcancer cells is their inability to differentiate normally (1-4). This defect may be partly reversible, and study of such models may help to elucidate the mechanism responsible for differentiation. The K-562 human leukemic cell line, which carries a Philadelphia chromosome and was established from a patient with chronic myelogenous leukemia in blast crisis (5), is such a model. This cell line is considered to consist of a multipotent hematopoietic stem cell, since it responds to inducers of erythroid differentiation (6-8), megakaryocytic differentiation (9, 10), and granulocytic differentiation (11,12). The cellular mechanisms responsible for these processes, however, are not known.cAMP, an intracellular regulatory agent, has long been considered to have a role in the control of cell proliferation and differentiation in a variety of cell types (13-16). cAMP exerts its effect in eukaryotes by binding to the regulatory subunits of either type I or type II cAMP-dependent protein kinase (17, 18). Differential expression ofthese protein kinase isozymes has been observed d...