Mechanisms of growth inhibition in keratinocytes by mercurio-substituted 4',5'-dihydropsoralens.

Psoralens, together with ultraviolet light A (PUVA), are used in the treatment of epidermal proliferative disorders. Although these compounds can enter cells and photo cross-link DNA, lipids and proteins, including a specific membrane receptor, are also potential targets for the psoralens. To better elucidate the site of action of the psoralens, we have synthesized a family of 5'-mercurio-substituted derivatives of 4',5'-dihydropsoralen. These compounds are identified by their heavy metal content and can be used as a model to deliver thiol reactive psoralen derivatives into keratinocytes. The 5'-mercuriopsoralen derivatives were found to be effective inhibitors of keratinocyte growth without photoactivation. The most active compound, 4,8-dimethyl-5'-iodomercuriomethyl-4',5'-dihydropsoralen (IC50=10 microM), was also a potent photosensitizer (IC50=0.3 microM). Depletion of keratinocyte GSH with buthionine sulfoximine markedly increased their sensitivity to this analog, both with and without UVA light. In contrast, N-acetyl-L-cysteine partially protected the cells from growth inhibition, indicating that a sulfhydryl-sensitive site is growth limiting and that this target can be photoactivated. Iodomercurio-4',5'-dihydropsoralen was found to form adducts with GSH and cysteine, which were not active without UVA light. Thus, these adducts may also contribute to the photosensitization reactions of the parent compound. Using plasmid DNA unwinding assays, iodomercurio-4',5'-dihydropsoralen was also found to modify DNA, an activity that increased following UVA light treatment. This suggests that DNA damage may contribute to the actions of these psoralens. Taken together, our data demonstrate that there are multiple sites of action for mercuriopsoralens. These compounds may prove useful for understanding the mechanisms of psoralen-induced growth inhibition in the skin.

Cell-impermeant pyridinium derivatives of psoralens as inhibitors of keratinocyte growth.

Psoralens such as 8-methoxypsoralen and 4,5',8-trimethylpsoralen (TMP) are used in photochemotherapy for the treatment of a variety of epidermal proliferative diseases. Sequential treatments of the skin with psoralens plus ultraviolet light in the range of 320-400 nm (UVA light), referred to as PUVA therapy, results in the suppression of abnormal keratinocyte growth. With the recognition that the psoralens are phototoxic and carcinogenic, presumably due to their ability to intercalate into DNA and photo cross-link pyrimidine bases following UVA light activation, it is clear that the development of biologically active analogs lacking this activity would be of significant therapeutic benefit. Towards this goal we have characterized active 4'- and 5'-pyridinium derivatives of 4',5'-dihydro-TMP (H2TMP), a psoralen analog that does not form DNA cross-links. These analogs, which are charged at physiological pH and cannot penetrate cells, are unique in that they retain biological activity as inhibitors of keratinocyte cell growth when activated by UVA light. However, they do not appear to cross-link or damage DNA as determined by plasmid DNA unwinding and nicking experiments, in intact cells using fluorescent analysis of DNA unwinding assays, and by thymidine uptake studies. Reverse transcription-polymerase chain reaction and western blotting demonstrated that, unlike TMP and H2TMP, when activated by UVA light, the pyridinium derivatives were not inhibitors of transcription since interferon-gamma-inducible nitric oxide synthase mRNA and protein in the keratinocytes were unaffected. Taken together, our data suggest that uptake of the compounds by the cells and DNA cross-link formation are not required for growth inhibition. These findings further support the model that the cell membrane is an important target for the psoralens.