The designing of DNA intercalating drugs with high DNA affinity in the series of ellipticine has led to a new antitumoral agent, 9-hydroxyellipticine, which has a high DNA affinity, a high activity on L 1210 mice leukemia, and a lack of toxicity at therapeutic dose. The possible correlations among chemical structure, DNA reactivity, and pharmacological activity of DNA intercalating drugs are discussed.Except for the hormonal products, which can prevent specific cells from proliferating, and drugs acting on the immunological system, most of the presently used anticancerous agents are cytotoxic compounds acting preferentially on tumor cells. Presently used antitumoral agents are already highly specific. For instance, it has been shown in an appropriate experimental model that a cancerous cell could be about a million times more sensitive to bis-chlorethylnitrosourea than a normal cell (1). Further progress in specificity seems, therefore, very difficult, especially since the basis at the molecular level of this specific toxicity is poorly understood. A rigorous approach in the design of new anticancerous compounds seems, therefore, an almost insuperable task. Nevertheless, the mechanism of the cytotoxicity property itself can in some cases be understood; that is mainly true for drugs acting on DNA structure or metabolism, among which are most of the anticancerous agents. If cytotoxic compounds could be rationally designed, one could hope to have better chance of finding among them anticancerous agents, since at the same time one could study the specificity of their action and try to understand it.Among products susceptible to such an approach are DNA intercalating drugs, because DNA intercalation is a rather well-understood mechanism and because it is almost the only case where the structure of what is thought to be the pharmacological receptor is known at atomic resolution (2). On the other hand, there belong already to this class of compounds some of the most powerful anticancerous agents, such as actinomycin D, daunomycin, and adriamycin. Our reasoning was, therefore, very simple. If DNA is the real receptor of these drugs and if we are able to design DNA intercalating compounds with the highest possible affinity for DNA, we would have a much better chance of finding active anticancerous drugs among these compounds. If this reasoning is correct, for this class of compounds high DNA reactivity is a condition necessary for anticancerous activity but not a condition that is sufficient; high DNA reactivity is necessary for conferring a potential cytotoxicity but is insufficient for conferring a specificity directed towards the cancerous cells.With these assumptions, the approach to be taken is clear. First, DNA intercalating drugs with increasing DNA affinities must be synthesized. Second, the correlation between the DNA reactivity and the pharmacological activity must be studied. The ability of drugs to intercalate in DNA is conditioned by their stereochemical parameters, such as size, shape, and planarity...