Before intelligent predictions can be made about the mode of action of chelating agents in the carcinogenic process or the design of specific chelating agents which may have anti-tumor properties, more information is needed about the properties of chelates themselves. Of prime importance may be data obtained by comparing the order of stability constants (KN) of suitable ligands for the biologically important cations; because of their importance in nucleic acid structure, both magnesium and calcium ions should be included in these studies. These thermodynamic constants should be obtained under physiological conditions of temperature, solvent, pH and ionic strength. In the Irving-Williams order of affinity of chelating agents for transition element ions, a plot of atomic number vs. log KN for biologically active compounds, such as glycine, ethylenediamine and salicyl-aldehyde, gives slopes that are similar but not necessarily parallel. Thus, the avidity of a ligand for a metal of lower atomic number may be greater than another ligand, whereas for a metal of a higher atomic number, the reverse may be true, as seen in Figure 1.

A comparison of KN values of two different ligands for the same cation will give limited information. The ratio of these two KN values will show relative affinities of each compound for that metal. The concentration of free cations in equilibrium with the chelates may be important, especially in a cell. Also to be considered must be the rate of ligand exchange. Methods are available for calculation of half-time for exchange. Note that many metals are more tightly bound to enzymes or biological material than to the chelate inhibitory agent. Even in these cases, an exchange could take place. The extent would be dependent upon both stability constants and rate of exchange.

Stability Of Two Series Of Chelates

The rate of formation of the chelate may also be important. Two different chelates may have almost the same stability constant, but the rates of formation of the complexes may be such that, for all practical purposes, only one complex will be formed in the finite time of investigation. In the case of hemoglobin, no measurable equilibrium may be possible between it and another iron chelating agent.

Activity of a chelate may be related to its solubility both in aqueous and lipid phases. Rates of excretion from the host and thus the biological half-life of the molecule may be important; a "good" drug may be eliminated before an adequate blood level is reached. Distribution of the agent to various tissues, even to the tumor, may be modified by the partition coefficient between the lipid and aqueous phases. The penetration of a chelate into cells may be affected by this latter property, too. The rate of transport of inorganic cations through cell membranes by chelates should also be considered. The ionic state of the molecule at physiological pH may also have some bearing on cell penetration. An example can be drawn from the field of microbiology. In a study of chelation and antifungal agents, it was found that both 8-hydroxyquinoline and its 5-sulfonic acid derivative have approximately the same affinity constant for some of the transition element ions. Only the 8-hydroxyquinoline penetrates the fungus cell and is a fungitoxic agent. The 5-sulfonic acid derivative is totally ineffective.

Fig. 1.

It is not possible to predict the nature of the bonds which will give the most stable chelate. Only qualitative statements can be made. A ligand with nitrogen may be expected to chelate copper rather than magnesium or calicum; these latter may be associated with oxygen. Zinc would be complexed with sulfur. Phosphates should bind magnesium. These statements do not imply mutual exclusion, for copper is strongly complexed by penicillamine, and quinoxaline dithiol is almost specific for nickel. Good studies are lacking in this field. More work on the relation of unit distances between functional groups of ligands to atomic radii of cations would be helpful. In which way are octahedrals distorted when complex ligands are attached? It is not possible to have two repelling groups attached to these geometric figures without distortion. The radii of most ions fall between the limits of 0.34 A to 1.65 A, and those metals associated with enzymes have limits between 0.78 A and 1.03 A. Can these data be used to explain ease of replacement of one essential cation by an abnormal one?