Proc. Natl. Acad. Sci. USA Vol. 83, pp. 4604-4606, July 1986 Biochemistry Phototoxicity of the tetracyclines: Photosensitized emission of singlet delta dioxygen (cutaneous phototoxicity/demeclocycline/minocycline)

TAYYABA HASAN* AND AHSAN U. KHANtt *Wellman Laboratories, Department of Dermatology, Harvard Medical School, Massachusetts General Hospital, Boston, MA 02114; tBiological Laboratories, Harvard University, Cambridge, MA 02138; and *Institute of Molecular Biophysics, Florida State University, Tallahassee, FL 32306 Communicated by Michael Kasha, January 23, 1986

ABSTRACT The spectroscopic observation of 1268-nm Table 1. Structure of tetracyclines emission of singlet oxygen photosensitized by tetracyclines in R2 R3 H 1 (CH3)2 oxygenated solutions at room temperature is reported. In the RI series demeclocycline, tetracycline, and minocycline, the effi- > ; S ~~~~OH ciency of singlet oxygen generation is found to parallel the clinical observation of relative frequency of phototoxicity of these antibiotics, suggesting singlet oxygen generation as the CONH2 origin of their phototoxicity. OH 0 HO 0 The tetracyclines are one of the most frequently prescribed Antibiotic R1 R2 R3 groups of antibiotics, deriving their bacteriostatic effect by Demeclocycline Cl H OH preventing the binding of the aminoacyl-tRNA to the Tetracycline H CH3 OH aminoacyl (A) site of the ribosome thus inhibiting protein Minocycline N(CH3)2 H H synthesis (1). A side effect associated with tetracycline therapy is cutaneous phototoxicity. The relative phototoxici- ties ofthe tetracyclines, estimated from clinical reports in the MATERIALS AND METHODS literature, clearly indicate that the chloro derivatives Table 1 the chlortetracycline and demeclocycline are the most potent shows structure ofthe three tetracyclines used in photosensitizers within the tetracycline family (2-7). The this investigation: tetracycline, demeclocycline, and mino- nature of the active species causing this phototoxicity is not cycline. Tetracycline hydrochloride, demeclocycline hydro- established, although activated oxygen species have been chloride, and minocycline hydrochloride (Sigma) were pre- postulated (8, 9). In this report we present the first direct pared in oxygen-saturated CC14 (Mallinckrodt; spectropho- evidence of tetracycline-photosensitized generation of sin- tometric grade) by first solubilizing the antibiotics in dimethyl glet delta oxygen in solution, identified by the 1268-nm sulfoxide (Me2SO) (Mallinckrodt; analytical reagent), then emission spectrum of 02 (1'A). We find that the yield of adding a fixed volume of the Me2SO solution to oxygenated photosensitized singlet oxygen follows the same order as the CC14. The final solution contained 0.6% (vol/vol) Me2SO in reported phototoxicity within the tetracycline family, sug- CCl4. The solutions containing the tetracyclines were of gesting that the basis for variation ofthe phototoxic potential comparable optical density, 0.070 ± 0.006 (1-cm cells) at 392 within the tetracycline series may lie in their relative effi- nm. The solutions were photoexcited on the long wavelength ciency of generating singlet oxygen upon photoexcitation. edge of the absorption bands of the tetracyclines. An exci- Consistent with clinical reports, we have previously shown tation band at 392 nm (389-404 nm) was isolated from a 75-W (9) that chlortetracycline and demeclocycline are also the Oriel xenon lamp by a combination of glass and liquid filters most effective photosensitizers in vitro, although the detailed [Corning filters CS 7-39, CS 3-75, and 5 cm of 4%5 (wt/vol) mechanism of tetracycline-induced phototoxicity is still not CuS04 and 4% (wt/wt) CoS04 in H20, cf. Fig. 1]. The clear. Based mainly on the ability of the tetracyclines to emission spectra were isolated (Coming filter CS7-56) and photooxidize the singlet oxygen scavengers dimethyl furan recorded with an ultrasensitive near-IR spectrometer based and limonene, Weibe and Moore (10) suggested that tetra- on an ADC 403L (Applied Detector, Fresno, CA) liquid cycline photosensitization in solution involved photooxygen- nitrogen-cooled germanium detector, Spex Minimate II, ation rather than a free radical process. However, the f/4.0 monochromator (Spex Industries, Metuchen, NJ) fitted specificity ofthe chemical scavengers used is not high enough with a 1.25-gm blazed grating, followed by a low noise to constitute definitive proof of the involvement of singlet PAR model Princeton oxygen. In cellular systems, based on the observation that in amplifier CR4, (E.G.&G. Applied vitro photosensitization was enhanced in 2H20 as compared Research, Princeton, NJ) lock-in amplifier (PAR model to H20, we (9) and others (8) have evoked singlet oxygen as HRB), followed by a Spex Datamate with digital storage and a possible reactive intermediate in tetracycline-induced pho- printout. The spectrometer is a modified version of the one totoxicity. This observation of 2H20 enhancement of photo- reported (12). sensitization, although supportive, is nevertheless indirect Quantum yields of photooxygenated singlet oxygen by tet- evidence of singlet oxygen involvement (11). In fact to our racyclines were estimated relative to the photoquantum yield of knowledge, prior to this study no direct demonstration ofthe singlet oxygen generated by aerated alkaline (10 mM KOH) tetracycline-photoinduced generation of singlet oxygen in ethanolic rose bengal at room temperature reported to be 0.75 solution or in cellular systems has been reported. (13). The areas under the peak of the 1268-nm singlet oxygen emission band photosensitized by demeclocycline in CC14/ The publication costs of this article were defrayed in part by page charge Me2SO solvent were compared to the same peak area in the payment. This article must therefore be hereby marked "advertisement" ethanolic rose bengal photosensitization, corrected for singlet in accordance with 18 U.S.C. §1734 solely to indicate this fact. oxygen lifetime in the solvent. The estimate of quantum yield

4604 Downloaded by guest on September 30, 2021 Biochemistry: Hasan and Khan Proc. Natl. Acad. Sci. USA 83 (1986) 4605 Table 2. Singlet oxygen photogeneration by the tetracyclines Relative Estimated efficiencies quantum yield Of '02 of '02 Clinical photogeneration* photogenerationt observations Demeclocycline 1.0 0.08 Most phototoxic Tetracycline 0.7 0.05 (6) Intermediate phototoxic Minocycline 0.0 0 Nonphototoxic *Singlet oxygen generation efficiency relative to demeclocycline. tSinglet oxygen quantum yield relative to alkaline rose bengal (ref. 13). tFrom refs. 2-7. for tetracycline and minocycline was made with respect to the photogeneration of singlet oxygen involves transfer of exci- demeclocycline value. The results are presented in Table 2. tation energy from an excited sensitizer molecule to ground state oxygen (31d. The dominant route of this electronic RESULTS energy transfer is via the triplet state of the photoexcited sensitizer molecule (18-20). In an halogen substituted series Electronic absorption spectra of the tetracyclines are char- of aromatic molecules the photopopulation ofthe triplet state acterized by major bands around 365 and 268 nm in solution follows the order H < Cl < Br < I, resulting from mixing of at room temperature and are presented in Fig. 1 for the three the electronic singlet and triplet states (21, 22). This expec- tetracyclines used in this investigation. Fig. 2 is the emission tation is consistent with the observed higher efficiency of spectral scan between 1200 nm and 1340 nm taken at a rate of 1 nm/sec with wide slits (8 mm x 8 mm) using the near IR singlet oxygen generation for demeclocycline compared to spectrometer. Two scans were accumulated, and the back- tetracycline. Minocycline, which is not phototoxic, did not ground containing solvent emission subtracted, showing the generate singlet oxygen in our experiments. Minocycline is 1268-nm singlet oxygen emission band of (0,0)1Ag 3-* 1g. The the only member of the series that has a dimethyl amine three traces shown are for demeclocycline, tetracycline, and substitution on the aromatic ring. Aromatic amines are well minocycline as labeled. The trace for minocycline shows no known singlet oxygen quenchers (23). emission in this spectral region. The estimated relative yields The site of tetracycline localization within the cell is not ofsinglet oxygen are presented in Table 2 and follow the trend clearly established. There are reports where tetracyclines demeclocycline > tetracycline > minocycline. have been shown to combine specifically with the mito- The self-sensitized photodestruction of the tetracyclines chondria of living cells both in tissue culture and in fresh also exhibits a molecular and structural dependence- preparations from various organs (24). Some evidence exists demeclocycline > tetracycline > minocycline-but the ab- that this is so in vivo (25) and that most of the intracellular solute rate for this photoreaction is slow. The 6-min exposure tetracycline in fact resides in the mitochondria (26) of to excitation radiation necessary to accumulate each spec- eukaryotic cells. A large proportion of the cellular oxidative trum does not significantly alter the results. processes occur in the mitochondria, and, ifthe mitochondria are considered potential subcellular targets for tetracycline- DISCUSSION Electronically excited singlet molecular oxygen is now rec- ognized as a distinct chemical reagent (14-17). The

(A

ct

._

._ u. 0 .0u, 4- ._

4._ r_

300 400 500 Wavelength, nm FIG. 1. Electronic absorption spectra of tetracyclines in 50% 1200 1270 1340 (vol/vol) Me2SO in MeOH between 254 nm and 550 nm at room Wavelength, nm temperature. For demeclocycline (DMCT), the molar extinction coefficient at 372 nm is 13.88 x 103 liter per mol per cm; for FIG. 2. Near IR emission of singlet oxygen in oxygenated [99.4% tetracycline (TC), the molar extinction coefficient at 366 nm is 14.15 CCL4/0.6% Me2SO (vol/vol)] solvent at room temperature, photo- X 103 liter per mol per cm; for minocycline (MC), the molar sensitized by demeclocycline (DMCT), tetracycline (TC), and mino- extinction coefficient at 350 nm is 11.39 x 103 liter per mol per cm. cycline (MC). The intensity is in arbitrary units. Downloaded by guest on September 30, 2021 4606 Biochemistry: Hasan and Khan Proc. NatL Acad Sci. USA 83 (1986) induced phototoxicity, the strong oxygen dependence and 10. Weibe, J. A. & Moore, D. E. (1977) J. Pharm. Sci. 66, possible singlet oxygen involvement are consistent. 186-189. In summary we report the first direct demonstration of 11. Merkel, P. B. & Kearns, D. R. (1972) J. Am. Chem. Soc. 94, tetracycline-photosensitized singlet oxygen generation in 1029-1030. 12. Khan, A. U. & Kasha, M. (1979) Proc. Natl. Acad. Sci. USA solution and find a qualitative correlation between the 102 76, 6047-6049. yield and the phototoxic potential of three tetracyclines. 13. Gadin, E., Lion, Y. & Van de Vorst, A. (1983) Photochem. More conclusive information is expected from in vitro and Photobiol. 37, 271-278. from in vivo experiments. 14. Khan, A. U. & Kasha, M. (1963) J. Chem. Phys. 39, 2105-2106. The authors acknowledge the generous hospitality of Professor J. 15. Foote, C. S. & Wexler, S. (1964) J. Am. Chem. Soc. 86, Woodland Hastings and Dr. Therese Wilson in giving them the 3879-3880. opportunity ofcarrying out the present research in their laboratories. 16. Corey, E. J. & Taylor, W. C. (1964) J. Am. Chem. Soc. 86, This work was supported by National Foundation for Cancer 3881-3882. Research, Bethesda, MD (A.U.K.), and Arthur 0. & Gullan M. 17. Schaap, A. P., ed. (1976) Singlet Molecular Oxygen, Bench- Wellman Foundation, Boston, MA (T.H.). mark Papers in Organic Chemistry (Dowden, Hutchinson and Ross, Stroudsburg, PA), Vol. 5. 18. Kawaoka, K., Khan, A. U. & Kearns, D. R. (1967) J. Chem. 1. Gale, E. F., Cundliffe, E., Reynolds, P. E., Richmond, M. H. Phys. 46, 1842-1853. & Waring, M. J. (1981) The Molecular Basis of Antibiotic 19. Kawaoka, K., Khan, A. U. & Kearns, D. R. (1967) J. Chem. Action (Wiley, New York), pp. 448-453. Phys. 47, 1883-1884. 2. Cullen, S. I., Catalano, P. M. & Helfman, R. J. (1966) Arch. 20. Khan, A. U. (1985) in Singlet Oxygen: Physical and Chemical Dermatol. 93, 77. Aspects, ed. Frimer, A. A. (CRC Press, Boca Raton, FL), Vol. 3. Schorr, W. F. & Monash, S. (1963) Arch. Dermatol. 88, 1, pp. 39-79. 134-138. 21. Kasha, M. (1950) Discuss. Faraday Soc. 9, 14-19. 4. Frost, P., Weinstein, G. D. & Gomez, E. C. (1971) J. Am. 22. McGlynn, S. P., Azumi, T. & Kinoshita, M. (1969) Molecular Med. Assoc. 216, 326-329. Spectroscopy of the Triplet State (Prentice-Hall, Englewood 5. Zuehlke, R. L. (1973) Arch. Dermatol. 108, 837-838. Cliffs, NJ), pp. 261-283. 6. Frank, S. B., Cohen, J. H. & Minkim, W. (1971) Arch. Der- 23. Monroe, B. M. (1985) in Singlet Oxygen: Physical and Chem- matol. 103, 520-521. ical Aspects, ed. Frimer, A. A. (CRC Press, Boca Raton, FL), 7. Blank, H., Cullen, S. I. & Catalano, P. M. (1968) Arch. Vol. 1, pp. 177-224. Dermatol. 97, 1-2. 24. Du Buy, H. G. & Showacre, J. L (1961) Science 133, 196-197. 8. Sandberg, S. O., Glette, J., Hopen, G. & Solberg, C. 0. (1984) 25. van den Bogert, C. & Kroon, A. M. (1981) Biochem. Photochem. Photobiol. 39, 43-48. Pharmacol. 30, 1706-1709. 9. Hasan, T., Kochevar, I. E., McAuliffe, D. J., Cooperman, 26. Babcock, D. F., Chen, J. J., Yip, B. P. & Lardy, H. A. (1979) B. S. & Abdulah, D. (1984) J. Invest. Dermatol. 83, 179-183. J. Biol. Chem. 254, 8117-8120. Downloaded by guest on September 30, 2021