Photosensitization by drugs*

Certain drugs are known to elicit photosensitivity side effects. A satisfactory understanding of the involved mechanistic aspects is necessary to anticipate the photosensitizing potential. We have used tiaprofenic acid (TPA), a photosensitizing nonsteroidal antiinflammatory drug, to illustrate the methodology followed to address this problem. After studying the photophysical and photochemical properties of TPA, the attention has been directed towards the reactivity of its lowest lying π−π* triplet with biomolecules. Photosensitized lipid peroxidation occurs by a mixed type I (radicals) and type II (singlet oxygen) mechanism. In the case of proteins, the photosensitized reactions include Tyr, Trp, and His photodegradation, protein–protein photocrosslinking and drug–protein photobinding. This involves direct quenching of the drug triplet by the amino acid residues (Tyr and Trp) or by oxygen, followed by singlet oxygen oxidation (His and Trp). With DNA, the studies have included comet assay, induction of single-strand breaks in supercoiled DNA, and reaction with 2'-deoxyguanosine and thymidine. Product studies, together with time-resolved measurements, have shown that the fastest reaction occurs with purine bases, by a mechanism involving both radical and singlet oxygen processes. The employed methodology can be of general use to investigate the mechanistic aspects of photosensitization by drugs. PHOTOSENSITIVITY SIDE EFFECTS The combined action of drugs and sunlight on patients can produce both desired and undesired effects [1]. Thus, PUVA-therapy (psoralenes plus UVA-radiation) has long been employed for the treatment of psoriasis, while porphyrins are currently being introduced for the photodynamic therapy (PDT) of cancer or other diseases. By contrast, there is also a significant number of reports indicating that a variety of drugs can elicit undesired side effects, such as phototoxicity, photoallergy, or photocarcinogenicity [2,3]. The photobiological risk associated with the use of drugs depends on environmental and individual factors (climate, height on the sea level, type of skin, etc.). On the other hand, the photosensitizing potential is enhanced in the case of topically administered drugs or when the field of application is dermatology or ophthalmology. Considering all these factors, in a number of cases it may be advisable to evaluate the photobiological risk of a new drug candidate before its introduction in the market [3]. The mechanistic approach to risk prediction In order to anticipate the appearance of photosensitivity side effects, a mechanistic understanding of the involved phenomena is necessary. Absorption of sunlight by drugs leads to their excited states. These can proceed further to afford drug-derived reactive intermediates or, under aerobic conditions, reactive oxygen species. Any of the above short-lived chemical entities may be able to interact with biological *Lecture presented at the XVIII IUPAC Symposium on Photochemistry, Dresden, Germany, 22–27 July 2000. Other presentations are published in this issue, pp. 395–548. substrates, ultimately producing photodamage. Thus, in a mechanistic approach the key questions are to determine which are the responsible light-absorbing chromophores, the intervening excited states, the reactive intermediates involved, the target biomolecules, and the reactions taking place [3]. A model group of drugs for photosensitization studies A number of drugs are capable of inducing photosensitivity disorders either after topical or systemic administration [2,3]. In particular, the nonsteroidal antiinflammatory 2-arylpropionic acids deserve special attention because they induce such disorders more frequently than other types of drugs [4–6]. Tiaprofenic acid (TPA) is a member of this family, that has been found to be one of the most potent photosensitizers in a multicenter photopatch test trial [5,6]. Its structure is that of 2-(4-[2-benzoyl]thienyl)propionic acid. In the present article, TPA has been chosen as a model compound to show the type of studies that can be undertaken in order to assess the photosensitizing potential of a given drug and to illustrate the methodological aspects of the problem. DRUG PHOTOPHYSICS AND PHOTOCHEMISTRY The absorption spectrum of TPA in neutral aqueous medium exhibits two intense bands with maxima at 266 and 314 nm and a weak tail extending up to 380 nm. The two main bands are shifted to the blue with decreasing solvent polarity [7]. The fluorescence of TPA appears at 420 nm and is very weak; its lifetime is lower than 0.5 ns. The phosphorence emission presents a maximum at 520 nm and is also very weak at room temperature. On the basis of the above absorption/emission spectra, complemented with theoretical calculations, the first excited singlet state is n–π*, and its energy is 81 kcal mol. By contrast, the lowest lying triplet state has a π–π* configuration, with an energy of 58 kcal mol. The second excited triplet is of n–π* nature and lies 10 kcal mol higher [7]. The photoreactivity of TPA is mediated by its lowest lying π–π* triplet state [7]. In aqueous medium, TPA is converted into decarboxytiaprofenic acid (DTPA), with a quantum yield of 0.25 at 25 °C [7,8]. The chromophore of TPA is maintained in the photoproduct; therefore, DTPA can mediate the same photosensitization processes elicited by the parent drug. Although DTPA is photostable in water, it undergoes photoreduction in hydrogen-donating solvents such as isopropanol. The resulting ketyltype radicals lead to hydrodimers as final products. On the basis of temperature-dependence studies and theoretical calculations, it has been concluded that both photodecarboxylation and photoreduction proceed from the triplet state; they require an activation energy between 7 and 10 kcal mol, which is essentially coincident with the energy difference between the two triplets. On the other hand, the lowest π–π* triplet states of TPA and DTPA are reactive towards phenols or indoles (used as models for the Tyr or Trp units of proteins). The process, which does not require thermal activation, occurs via electron transfer from the donor to the (D)TPA π–π* triplet, followed by proton transfer. The result is again formation of the same ketyl radicals generated after hydrogen abstraction [7]. The intermediates involved in the photochemistry of TPA and DTPA have been characterized by nanosecond laser flash photolysis [9]. Photoexcitation of TPA at 355 nm in aqueous medium leads to the lowest π–π* triplet with a very high efficiency (ca. 0.9). The triplet is detected by its transient absorption, with maxima at 380 and 590 nm. Its deactivation occurs in the microsecond timescale (lifetime 0.8 μs) and is dominated by a thermally activated spin-allowed process (activation energy barrier ca. 10 kcal mol). At neutral pH, there is an adiabatic loss of carbon dioxide, leading to a triplet biradical anion. This intermediate undergoes intersystem crossing to give a long-lived decarboxylated carbanion, which is finally protonated. In the case of DTPA, a similar triplet is also detected upon laser flash photolysis [9]. This species is essentially unreactive in aqueous medium, and hence its lifetime (6 μs) is markedly higher than that of triplet TPA. However, hydrogen abstraction in isopropanol is demonstrated by the diminished triplet lifetime (3.2 μs) and detection of the ketyl radical, which absorbs at 350 and 390 nm and decays by a second-order kinetics due to dimerization. M. A. MIRANDA © 2001 IUPAC, Pure and Applied Chemistry 73, 481–486 482