- Email: [email protected]
Time-Course of Hypericin Phototoxicity and Effect on Mitochondrial Energies in EMT6 Mouse Mammary Carcinoma Cells
Free Radical Biology & Medicine, Vol. 25, No. 2, pp. 144 –152, 1998 Copyright © 1998 Elsevier Science Inc. Printed in the USA. All rights reserved 0891-5849/98 $19.00 1 .00
PII S0891-5849(98)00052-5
Original Contribution TIME-COURSE OF HYPERICIN PHOTOTOXICITY AND EFFECT ON MITOCHONDRIAL ENERGIES IN EMT6 MOUSE MAMMARY CARCINOMA CELLS SANDRA A. S. JOHNSON, AMY E. DALTON,
and
RONALD S. PARDINI
Cancer Research Laboratory and the Natural Products Laboratory, Department of Biochemistry, University of Nevada, Reno, NV, USA (Received 18 July 1997; Revised 22 December 1997; Accepted 6 January 1998)
Abstract—Photoactivated hypericin produces singlet oxygen and superoxide anion radical; however, the intracellular events contributing to toxicity are unknown. Clonogenic assays of oxygen-dependent hypericin phototoxicity to EMT6 cells have previously shown that 0.5 mM hypericin 1 1.5 J cm22 fluorescent light is non-toxic and that 1.0 mM hypericin 1 1.5 J cm22 fluorescent light produces LD40 toxicity. Intracellular events leading to toxicity were revealed at these doses. Lactate dehydrogenase leakage was elevated for both 0.5 mM and 1.0 mM hypericin 1 light immediately following irradiation. While values eventually returned to control levels for 0.5 mM hypericin 1 light, leakage increased over time for 1.0 mM hypericin indicating reversible and irreversible toxicity, respectively. Increases in lipid and protein oxidation were measured immediately following irradiation; however, these parameters return to control levels within 0.5 h for both doses. Both total cellular ATP levels and cellular respiration were depressed by approximately 50% of control values for 1.0 mM hypericin 1 light. These values were unchanged for 0.5 mM hypericin 1 light. Along with previously reported data demonstrating that light-activated hypericin can inhibit mitochondial succinoxidase in beef heart mitochondria in vitro, these data support oxidative stress-initiated mitochondrial damage as a key target in hypericin phototoxicity. © 1998 Elsevier Science Inc. Keywords—Anticancer agents, Antiviral agents, Hypericin, Photosensitizers, Photodynamic therapy, Oxidative damage, Mitochondria, Free radical
INTRODUCTION
these reports, the cellular mechanism of phototoxicity is not well defined. The mitochondria has been proposed as a potential target of photodynamic action. Hematoporphryin derivative (HPD) photosensitization has been shown to inhibit mitochondrial enzymes involved in oxidative phosphorylation in vitro, including the proton translocating ATPase, cytochrome C oxidase, and succinate dehydrogenase.8,9,10 Later, it was demonstrated that cellular ATP levels decreased in a drug and a light dose-dependent manner with HPD PDT and this decrease was directly related to impairment of mitochondrial function.11 More recently, photoactivated hypericin has been shown to inhibit isolated bovine heart mitochondrial succinoxidase in a drug and a light dose-dependent manner.2 These results were further supported in a study with isolated rat liver mitochondria in which membrane potential and succinoxidase activity were decreased along with respiratory control ratio and ADP/O in the presence of hypericin and light.12
Hypericin is emerging as a good candidate sensitizer for the photodynamic therapy (PDT) of neoplastic diseases. The phototoxic effect of hypericin is thought be oxygendependent, presumably via a type II mechanism involving singlet oxygen.1 Recent data are strongly supportive of this hypothesis,2,3,4 although supporting roles for superoxide anion radical, hydrogen peroxide, and hydroxyl radical have been suggested.4,5 Hypericin generates singlet oxygen at a high quantum yield of 0.746 in comparison with other photosensitizers for which the quantum yield ranges from 0.2– 0.6.7 These reactive oxygen species are assumed to be the agents of hypericin toxicity in EMT6 cells in cell culture since hypericin demonstrated no toxicity under hypoxic or dark conditions.3 Beyond Address correspondence to: Ronald S. Pardini, University of Nevada, Department of Biochemistry MS 330, Reno, NV 89557; Tel: 702-784-4107; Fax: 702-784-1419; E-Mail: [email protected]. 144
Time-course of hypericin phototoxicity
In the present study we establish a time frame to investigate the mode of hypericin phototoxicity for EMT6 mouse mammary carcinoma cells in vitro. The previously reported non-toxic and toxic doses, 0.5 mM and 1.0 mM hypericin, respectively, with 1.5 J cm22 fluorescent light (cell survival: 1.081 6 0.108 and 0.609 6 0.032 respectively)3 were utilized to elucidate the early events of hypericin phototoxicity in EMT6 cells. Lactate dehydrogenase leakage assay was used to measure the level of membrane damage over a 3 h period beginning at the end of the irradiation period. Indications of oxidative stress were measured by lipid peroxidation and protein oxidation assays within this period as well. To further investigate the potential role of mitochondrial impairment in the mechanism of hypericin phototoxicity, we examined the effect of photoactivated hypericin on total cellular ATP concentrations and total cellular respiration in EMT6 cells. METHODS
Chemicals Hypericin was obtained from Atomergic Chemetals (Farmingdale, NY). Powdered Waymouth’s media, neonatal clostrum free calf serum, L-glutamine, trypsin, penicillin-G/streptomycin, dimethylsulfoxide (DMSO), and ATP assay kit were obtained from Sigma Chemical Co. (St. Louis, MO). Dulbecco’s phosphate buffered saline (DPBS) and enzyme grade sodium bicarbonate were obtained from Fisher Scientific (Fair Lawn, NJ). All other chemicals were of analytical grade. EMT6 cell culture The characteristics of the EMT6 mouse mammary carcinoma cell line and techniques used for maintaining cultures have been previously reported.13,14 Cells were grown in flasks as monolayers at 37°C in a saturated humidified environment consisting of 5% CO2 and 95% air. Cells were grown in complete media which is Waymouth’s medium supplemented with 15% neonatal colostrum free calf serum 118 units/ml penicillin G, and 119 mg/ml streptomycin. Tissue culture flasks were seeded with 1 3 105 cells/ml of complete media and all tests were conducted on cells in exponential growth phase. Hypericin exposure Cells were washed thoroughly with DPBS and complete media was replaced with fresh complete media prior to exposing anchored cells to hypericin or to DMSO solvent treatment for 1 h. The final concentration
145
of DMSO in each flask was 0.4% (v/v). At the end of hypericin exposure, cells were washed thoroughly with DPBS. Irradiation All procedures, with the exception of irradiation, were carried out in subdued light (,0.010 mW cm22). Anchored cells in flasks were placed on a 3 mm thick glass 1 cm above 8 Philips cool white 24T12 high output fluorescent bulbs. The light dose was measured with an IL 1700 radiometer (International Light; Newburyport, MA) by integrating the energy signal in mW cm22 over the entire period of irradiation (10.31 6 1.01 min). Over the irradiation period, the light intensity decreased from an initial intensity of 2.74 6 0.17 mW cm22 to a final intensity of 2.14 6 0.21 mW cm22. The light source was extinguished when the total light dose reached 1.5 J cm22. Fluorescent light intensity was similar when measured through plastic flasks used for tissue culture. Temperature measurements recorded in tissue culture flasks during irradiation never exceeded 37°C and decreased to no lower than 31°C by the end of irradiation. Cells in flasks serving as dark controls were placed under a cardboard box at room temperature during the irradiation period. Clonogenic cell survival studies with conditions of irradiation (no hypericin) described here have been previously investigated and demonstrated no substantial alteration in cell survival compared to dark control.3 Therefore, it is assumed that no toxicity occurs in EMT6 mouse mammary carcinoma cells due to irradiation alone. Post-irradiation period incubation Cells either were assayed immediately following the irradiation period (0 h) or were incubated for 0.5 h, 1 h, or 3 h beginning when the complete media was replaced in all flasks. Time between extinguishing the light source and replacing the complete media in all flasks was not more than 5 min. At the end of post-irradiation incubation time, media was decanted off and cells were removed from the flasks manually by scraping. Samples were placed on ice for the remainder of testing. Lactate dehydrogenase (LDH) leakage Following hypericin exposure and rinsing, complete media was replaced and cells were irradiated as described above. Cells were handled as described for the post-irradiation period above except that they remained in the complete media in which they were placed following the hypericin incubation. Immediately following
146
S. A. S. JOHNSON et al.
irradiation, or at the end of the post-irradiation incubation, cells were scraped within the flask and mixed thoroughly with the complete media within the flask. A known volume of media/cell mixture was removed and centrifuged at 1,000 3 g for 5 min. Media supernatant (0.5 ml) was removed to measure the levels of LDH leaked. The balance of the mixture was then sonicated to measure total LDH activity. LDH activity was measured as described by Kornberg.15 Activity of complete media alone was measured and subtracted from the activities of the media supernatant plus the sonicated mixture prior to calculating total percentage of LDH activated leaked. Lipid peroxidation and protein oxidation Following the appropriate incubation time, cells were removed by scraping then were centrifuged at 1,000 3 g for 5 min. Remaining supernatant was decanted and replaced with a lysing buffer (10 mM Tris, 1 mM EDTA, pH 7.5, 2% Triton X-100) then vortexed vigorously and sonicated. Complete preparation of cell extracts took no longer than 20 min from the time of cell removal to sonication of the cell mixture. Thiobarbituric reactive species (TBARS) were measured as an indication of lipid peroxidation as described by Esterbauer and Cheeseman.16 Whole cell homogenates were mixed with 20% (w/v) trichloroacetic acid (half of the original volume) and 0.67% (w/v) thiobarbituric acid (equal amount of the original volume) then heated in a boiling water bath for 10 min. Precipitate was centrifuged out at 10,000 3 g for 10 min. Thiobarbituric reactive species in the supernatant were then measured spectrophotometrically at 535 nm (« 5 1.56 3 105 M21 cm21). Absorbances of lysing buffer controls were subtracted from test sample absorbances. Protein carbonyl groups were measured as an indication of protein oxidation according to the method described by Levine, et al.17 Schiff’s bases were also measured by this procedure. DNA was removed from the fraction by streptomycin precipitation. Remaining proteins were precipitated and the pellet was mixed with 10 mM 3,4-dinitrophenyl hydrazine (DNPH) for 1 h. Excess DNPH was removed with a 1:1 ethanol:ethyl acetate wash mixture and the pellet was re-dissolved in 0.6 M guanidine in 20 mM potassium phosphate buffer, pH 6.3. Solution absorbance was then measured at 390 nm («360 –390 nm 5 22,000 M21 cm21) to determine the presence of protein carbonyl groups. Control sample absorbances without DNPH were subtracted from test absorbances. ATP assay Immediately following irradiation, cells were washed 3 times with DPBS, removed from the flasks with 0.05%
trypsin diluted in DPBS, and then counted with a Coulter counter (Coulter Scientific, Hialeah, FL). Cells were centrifuged at 1,000 3 g for 5 min, and the remaining media supernatant was removed. The cell pellet was then rinsed with DPBS and re-centrifuged at 1,000 3 g for 5 min. DPBS supernatant was removed, and 500 ml of 67 mM Tris, 33.3 mM magnesium sulfate, 6.7 mM EDTA, pH 7.8 was added. The cell mixture was then sonicated and precipitated with 100 ml cold 20% (w/v) trichloroacetic acid. The mixture was neutralized with 6 M potassium hydroxide and centrifuged at 10,000 3 g for 3 min. Total cellular ATP for each test was measured by bioluminescent assay (Sigma, FL-AA) using Na2ATP in a standard curve on a Lab System Luminoskan Luminometer. Test sample ATP concentrations were compared to ATP levels in control cells receiving no treatment during the same isolation to get percent of control values. Cellular respiration Immediately following irradiation, cells were washed 3 times with DPBS, removed from the flasks with 0.05% trypsin diluted in DPBS, and then counted with a Coulter counter (Coulter Scientific, Hialeah FL). Cells were centrifuged at 1,000 3 g for 5 min and remaining media supernatant was removed. Cells were diluted to a final concentration of 2.5 3 106 cells/ml in complete media. Complete Waymouth’s media contains a final concentration of 5 g/l glucose and essential amino acids. Other than this, no other substrates of glycolysis, electron transport, or oxidative phosphorylation were supplemented during the measurement of oxygen consumption. Cells were kept on ice until oxygen consumption was measured using a Gilson water jacketed reaction chamber (Middleton, WI) set at 37°C and equipped with a Yellow Springs Instruments model 53 oxygen monitor (Yellow Springs, OH) and a Clark electrode. Complete media oxygen consumption was measured prior to addition of cells and was not significant. Cellular oxygen consumption was measured for a minimum of 5 min. Protein assays Protein concentrations were determined using the BCA kit (Pierce, Rockford, IL). Statistical analysis Data were analyzed by analysis of variance using SAS Systems for Windows release 6.11 TS Level 0040 computer software (SAS Institute Inc., Cary, NC). If the analysis exhibited a significant difference, data were
Time-course of hypericin phototoxicity
147
Fig. 1. Hypericin toxicity towards EMT6 mouse mammary carcinoma cells. Percent of LDH leaked measured immediately following irradiation with 1.5 J cm22 fluorescent light. Cells were incubated with hypericin 1 h prior to irradiation. Statistical differences based on Tukey–Kramer least square means comparisons: alight vs. dark control; bhypericin vs. DMSO control.
further analyzed by Tukey–Kramer least square means comparisons to discern significant differences among the means ( p , .05). RESULTS AND DISCUSSION
Phototoxicity to EMT6 mouse mammary carcinoma cells has been established.3 The survival curve generated in that study showed that tolerance for hypericin phototoxicity extended up to 0.5 mM hypericin with light and then diminished sharply thereafter. LDH leakage was measured in EMT6 cells immediately following irradiation to determine the amount of membrane damage that occurred during the photoactivation period (Fig. 1). The effect was dose and light-dependent with the highest dose measured, 2.5 mM hypericin 1 light, sustaining enough membrane damage to demonstrate 91.4% of total activity leaked indicating almost instantaneous cytotoxicity for these cells. The levels of leakage experienced by cells treated with 0.5 mM hypericin 1 light and 1.0 mM 1 light immediately following irradiation were elevated over their dark controls but were not statistically different from one another (Fig. 1 and 2, p , 0.99). We know from the previous clonogenic assays that 0.5 mM
hypericin 1 light is not toxic to EMT6 cells while 1.0 mM hypericin 1 light is marginally toxic.3 This implies that photoactivated hypericin elicited enough damage at the non-toxic dose to allow elevated membrane leakage but that the dose was not strong enough to evoke eventual toxicity while the higher dose, experiencing similar levels of initial membrane damage, was unable to recover. Given this information, we decided that these two doses could be targeted to dissociate key events leading to eventual hypericin phototoxicity in EMT6 mouse mammary carcinoma cells. The equivalent levels of initial LDH leakage between the non-toxic and marginally toxic doses led us to focus on later events in the time-course of eventual hypericin toxicity. Following the irradiation period, cells were incubated for 0.5, 1, and 3 h. Consistent with the clonogenic data, total activity leakage increased over a time period of 3 h for cells exposed to 1.0 mM hypericin 1 light (Fig. 2). In contrast, cells exposed to 0.5 mM hypericin 1 light demonstrated no further indication of membrane damage and intracellular LDH levels were restored back to the level of the controls (Fig. 2). The effect of phototoxicity with hypericin is oxygendependent.3 The fact that hypericin is known to produce
148
S. A. S. JOHNSON et al.
Fig. 2. LDH leakage following treatment with hypericin 6 1.5 J cm22 fluorescent light from EMT6 mouse mammary carcinoma cells. Cells were prepared for assay immediately following irradiation (0 h) or allowed to incubate in complete media following the irradiation period (0.5 h, 1 h, or 3 h). Values are means 6 SEM of 5–35 independent determinations. Statistical differences based on Tukey–Kramer least square means comparisons: alight vs. dark control; bhypericin vs. DMSO control; cadditional relevant statistical differences: 1.0 mM hypericin 1 light: 0 h vs. 1.0 mM hypericin 1 light: 0.5 h, 1 h, 3 h.
high quantum yields of singlet oxygen6 along with superoxide anion radical4,5 suggests a mechanism of phototoxicity in which the generation of reactive oxygen species causes oxidative damage which must be demonstrated cellularly. Evidence of oxidative damage was investigated by measuring indices of lipid peroxidation and protein oxidation on the same time-course basis used previously. Both TBARS and protein carbonyls were elevated immediately following irradiation for both 1.0 mM and 0.5 mM hypericin 1 light but returned to control levels within 0.5 h (Fig. 3). While lipid peroxidation levels were higher for 1.0 mM hypericin 1 light than for 0.5 mM hypericin 1 light, protein oxidation was equivalent for both doses. These results are supportive of previous studies which have demonstrated hypericin’s ability to induce lipid peroxidation18,19 and further establishes that oxidative damage occurs at a cellular level as well. The fact that these general measures of oxidative damage are absent after 0.5 h of the post-irradiation period suggests that EMT6 cells were able to repair this
damage for both doses. Yet despite this, 1.0 mM hypericin 1 light is toxic to EMT6 cells and membrane damage is evident (Fig. 1, 2). Hypericin toxicity has been consistently shown to involve reactive oxygen species as antioxidants and free radical scavengers, in addition to hypoxic conditions, have all been shown to protect cells from photoactivated hypericin.3,20 This implies that reactive oxygen species generated by photoactivated hypericin may have more discrete targets which are unable to be repaired following photo-oxidative damage at 1.0 mM hypericin 1 light in EMT6 cells. It is also possible that damage to critical targets arises from the initial oxidative damage. Mitochondria are considered a primary target of phototherapy,21,22 and hypericin association with cellular membranes has been observed.3,23 As seen in Figure 4A, total cellular ATP levels in EMT6 mouse mammary carcinoma cells were decreased to 53.6 6 2.4% of controls by 1.0 mM hypericin 1 light. This decrease is light-dependent and is consistent with previous cytotox-
Time-course of hypericin phototoxicity
Fig. 3. Lipid peroxidation (panel A) and protein oxidation (panel B) following treatment with hypericin 6 1.5 J cm22 fluorescent light in EMT6 mouse mammary carcinoma cells. Lipid peroxidation was measured as thiobarbituric reactive species absorbing at 532 nm. Protein oxidation was measured as protein carbonyls reacting with 2,4-dinitrophenyl hydrazine. Cells were prepared for assay immediately following irradiation (0 h) or allowed to incubate in complete media following the irradiation period (0.5 h, 1 h, or 3 h). Values are means 6 SEM of 3–9 independent determinations. Statistical differences based on Tukey–Kramer least square means comparisons: alight vs. dark control; bhypericin vs. DMSO control; cadditional relevant statistical differences: Lipid peroxidation: 1.0 mM hypericin 1 light: 0 h vs. 1.0 mM hypericin 1 light: 0.5 h, 1 h, 3 h; 0.5 mM hypericin 1 light: 0 h vs. 0.5 mM hypericin 1 light: 0.5 h, 1 h, 3 h. Protein oxidation: 1.0 mM hypericin 1 light: 0 h vs. 1.0 mM hypericin 1 light: 0.5 h, 1 h, 3 h; 0.5 mM hypericin 1 light: 0 h vs. 0.5 mM hypercin 1 light: 0.5 h, 1 h, 3 h; DMSO 1 light: 1 h vs. DMSO 1 light: 3 h.
149
150
S. A. S. JOHNSON et al.
Fig. 4. Total cellular ATP (panel A) and total cellular respiration (panel B) in EMT6 mouse mammary carcinoma cells following treatment with hypericin 6 1.5 J cm22 fluorescent light. Cells were incubated with hypericin 1 h prior to irradiation. ATP data is present as percent control (means 6 SEM, n 5 5–14) where tested ATP levels were compared to total ATP extracted from control cells (no treatment) during the same extraction. For cellular respiration, cells were suspended in complete Waymouth’s media containing a final concentration of 5 g/l glucose and essential amino acids. Other than this, no other substrates of glycolysis, electron transport, or oxidative phosphorylation were added during the measurement of oxygen consumption. Cellular respiration (means 6 STD, n 5 5–9) was measured as oxygen consumption with a Gilson water jacketed reaction chamber equipped with a Yellow Springs Instrument model 53 oxygen monitor and a Clark electrode. Statistical differences based on Tukey–Kramer least square means comparisons, alight vs. dark control; bhypericin vs. DMSO control.
icity data. In addition, 1.0 mM hypericin 1 light proportionally inhibited total cellular respiration (Fig. 4B) to about half of control respiration rates. Cells treated with 1.0 mM hypericin consumed only 0.00076 6 0.00001 matoms O/min per 106 cells while cells treated with DMSO 1 light consumed 0.00143 6 0.00013 matoms
O/min per 106 cells. In contrast, non-toxic 0.5 mM hypericin with and without light had no effect on total cellular ATP levels and total cellular respiration (Fig. 4A, B). These observations are supportive of previous data which demonstrated hypericin’s ability to impair mitochondrial function in vitro in a drug and light-dose
Time-course of hypericin phototoxicity
dependent manner.2,12 More recently, release of mitochondrial hexokinase activity from the mitochondria as well as inhibition of this activity has been induced by photoactivated hypericin in human glioma cells.24 This effect was accompanied by a hypericin-induced decrease in intracellular pH in addition to decreases in ATP and glucose-6-phosphate. While the hexokinase release from mitochondria was attributed to hypericin-dependent intracellular acidification, hexokinase inhibition was oxygen-dependent.24,25 The dual effects of oxidative damage and change in intracellular pH elicited by photoactivated hypericin could be the precise combination which makes this drug, “one of the most powerful photodynamic agents in nature.”26 In combination with the current observations, it is reasonable to extend the possible role of mitochondria as a target of oxygen-dependent hypericin phototoxicity from a sub-cellular to a cellular level. CONCLUSIONS
Photoactivated hypericin initiates a series of oxidative stress events at a cellular level. At the non-toxic and toxic doses tested, two indicators of oxidative stress, lipid peroxidation and protein oxidation, were initially elevated then returned to control levels for both doses in EMT6 cells. Membrane damage is evident initially for nontoxic 0.5 mM hypericin 1 light and toxic 1.0 mM hypericin 1 light. Cells exposed to 0.5 mM hypericin 1 light repaired damage within 3 h, consistent with clonogenic toxicity data. Membrane damage for 1.0 mM hypericin 1 light persisted for the duration of the experiment. These doses will be useful in differentiating the key events that ultimately end in cellular toxicity. One possible discrete target of hypericin phototoxicity may be mitochondria. At 1.0 mM hypericin 1 light, cellular respiration was depressed by 47% compared to DMSO 1 light controls. Similarly, cellular ATP concentrations were depressed by 47% compared to untreated controls by the same hypericin treatment. These findings suggest that photoactivated hypericin interact at the mitochondrial level in EMT6 cells. This conclusion is supported by our recent report that MnSOD was initially elevated and subsequently depressed over a 3 h time period following treatment with 1.0 mM hypericin 1 light.27 Further investigation of photoactivated hypericin’s effect on glycolysis, electron transport, and oxidative phosphorylation in whole cells will help further elucidate the role of mitochondria in hypericin’s mechanism of cellular toxicity. Acknowledgements—Funding for these studies was provided in part by The Grand Chapter of Nevada, Order of the Eastern Star and the Howard Hughes Medical Institute Undergraduate Biological Sciences Education Program, Grant #71192-516901. We are grateful to Laurent
151
Miccoli for helpful discussion. We thank Megan Probst for technical assistance and Dr. Chris Pritsos for providing EMT6 cells for our culture experiments.
REFERENCES 1. Duran, N.; Song, P.-S. Yearly Review: Hypericin and its photodynamic action. Photochem. Photobiol. 43:677– 680; 1986. 2. Thomas, C.; MacGill, R. S.; Miller, G. C.; Pardini, R. S. Photoactivation of hypericin generates singlet oxygen in mitochondria and inhibits succinoxidase. Photochem. Photobiol. 55:47–53; 1992. 3. Thomas, C,; Pardini, R. S. Oxygen dependence of hypericin induced phototoxicity to EMT6 mouse mammary carcinoma cells. Photochem. Photobiol. 55:831– 837; 1992. 4. Diwu, Z.; Lown, J. W. Photosensitization with anticancer agents 17. EPR studies with photodynamic action of hypericin: formation of semiquinone radical and activated oxygen species on illumination. Free Radical Biol. Med. 14:209 –215; 1993. 5. Hadjur, C.; Jeunet, A.; Jardon, P. Photosensitization by hypericin: electron spin resonance (ESR) evidence for the formation of singlet oxygen and superoxide anion radicals in an in vitro model. J. Photochem. Photobiol. B.: Biol. 26:67–74; 1994. 6. Jardon, P.; Lazortchak, N.; Gautron R. Formation d’oxygene singluet 1Dg photosensibilisee par l’hypericine. Caracterisation et etude du mecanisme par spectroscopie laser. J. Chim. Phys. 84: 1143–1145; 1987. 7. Pass, H. I. Photodynamic therapy in oncology: mechanism and clinical use. J. Natl. Cancer Inst. 85:443– 456; 1993. 8. Gibson, S. L.; Hilf, R. Photosensitization of mitochondrial cytochrome C oxidase by hematoporphyrin derivative and related porphyrins in vitro and in vivo. Cancer Res. 43:4191– 4197; 1983. 9. Hilf, R.; Small, D. B.; Murant, R.; Leaky, P. B.; Gibson, S. Hematoporphyrin derivative-induced photosensitivity of mitochondrial succinate dehydrogenase and selected cytosolic enzymes of R3230AC mammary adenocarcinomas of rats. Cancer Res. 44:1483–1488; 1984. 10. Perlin, D. S.; Murant, R. S.; Gibson, S. L.; Hilf, R. Effect of photosensitization by hematoporphyrin derivative on mitochondrial adenosine triphosphatase-mediated proton transport and membrane integrity of R3230AC mammary adenocarcinoma. Cancer Res. 45:653– 658; 1985. 11. Hilf, R.; Murant, R. S.; Narayan, U.; Gibson, S. L. Relationship of mitochondrial function and cellular adenosine triphosphate levels to hematoporphyrin derivative-induced photosensitization in R3230AC mammary tumors. Cancer Res. 46:211–217; 1986. 12. Utsumi, T.; Okuma, M.; Kanno, T.; Takehara, Y.; Yoshioka, T.; Fujita, Y,; Horton, A. A.; Utsumi, K. Effect of the antiretroviral agent hypericin on rat liver mitochondria. Biochem. Pharmacol. 50:655– 662; 1995. 13. Rockwell, S. C.; Kallman, R. F.; Fajardo, L. F. Characteristics of a serially transplanted mouse mammary tumor and its tissueculture-adapted derivative. J. Natl. Cancer Inst. 49:735–747; 1972. 14. Rockwell, S. C. In vivo-in vitro tumor systems: new models for studying the response of tumors to therapy. Lab. Anim. Sci. 27: 831– 851; 1977. 15. Kornberg, A. Lactate dehydrogenase of muscle. Methods Enzymol. 1:441– 443; 1955. 16. Esterbauer, H.; Cheeseman, K. H. Determination of aldehydic lipid peroxidation products: malondialdehyde and 4-hydroxynonenal. Methods Enzymol. 186:407– 421; 1990. 17. Levine, R. L.; Garland, D.; Oliver, C. N.; Amici, A.; Climent, I.; Lenz, A.-G.; Ahn, B.; Shaltiel, S.; Stadtman, E. R. Determination of carbonyl content in oxidatively modified proteins. Methods Enzymol. 186:464 – 478; 1990. 18. Hadjur, C.; Richard, M.-J.; Parat, M.-O.; Jardon, P.; Favier, A. Photodynamic effects of hypericin on lipid peroxidation and antioxidant status in melanoma cells. Photochem. Photobiol. 64:375– 381; 1996. 19. Knox, J. P.; Dodge, A. D. Isolation and activity of the photodynamic pigment hypericin. Plant Cell Environ. 8:19 –25; 1985.
152
S. A. S. JOHNSON et al.
20. Hadjur, C.; Richard, M. J.; Parat, M. O.; Favier, A.; Jardon, P. Photodynamically induced cytotoxicity of hypericin on human fibroblast cell line MRC5. J. Photochem. Photobiol. B: Biol. 27: 139 –146; 1995. 21. Hilf, R.; Gibson, S. L.; Penney, D. P.; Checkler, T. L.; Bryant, R. C.; Mitochondria in cancer cells as targets of photodynamic therapy. Proc. Soc. Photo-Opt Instrum. Eng. 847:2–10; 1989. 22. Salet, C.; Moreno, G. New trends in photobiology (Invited Review): photosensitization of mitochondria. Molecular and cellular aspects. J. Photochem. Photobiol. B: Biol. 5:133–150; 1990. 23. Miskovsky, P.; Sureau, F.; Chinsky, L.; Turpin, P.-Y. Subcellular distribution of hypericin in human cancer cells. Photochem. Photobiol. 62:546 –549; 1995. 24. Miccoli, L.; Sureau, F.; Dutrillaux, B.; Poupon, M.-F. Hypericin disrupts the hexokinase mitochondria-binding and affects the pHdependent metabolism of human gliomas. Proc. Am. Assoc. Cancer Res. 38:598; 1997. 25. Miccoli, L.; Oudard, S.; Sureau, F.; Dutrillaux, B.; Poupon, M.-F. Intracellular pH governs the subcellular distribution of hexokinase in a glioma cell line. Biochem. J. 313:957–962; 1996.
26. Walker, E. B.; Lee, T. Y.; Song, P.-S. Spectroscopic characterization of the stentor photoreceptor. Biochim. Biophys. Acta 587:129 – 144; 1979. 27. Johnson, S. A. S.; Pardini, R. S. Antioxidant enzyme response to hypericin in EMT6 mouse mammary carcinoma cells. Free Radical Biol. Med. 24:817– 826; 1998.
ABBREVIATIONS
DNPH— 4-dinitrophenyl hydrazine DMSO— dimethyl sulfoxide DPBS—Dulbecco’s phosphate buffered saline HPD— hematoporphyrin derivative LDH—lactate dehydrogenase PDT—photodynamic therapy TBARS—thiobarbituric reactive species
PII S0891-5849(98)00052-5
Original Contribution TIME-COURSE OF HYPERICIN PHOTOTOXICITY AND EFFECT ON MITOCHONDRIAL ENERGIES IN EMT6 MOUSE MAMMARY CARCINOMA CELLS SANDRA A. S. JOHNSON, AMY E. DALTON,
and
RONALD S. PARDINI
Cancer Research Laboratory and the Natural Products Laboratory, Department of Biochemistry, University of Nevada, Reno, NV, USA (Received 18 July 1997; Revised 22 December 1997; Accepted 6 January 1998)
Abstract—Photoactivated hypericin produces singlet oxygen and superoxide anion radical; however, the intracellular events contributing to toxicity are unknown. Clonogenic assays of oxygen-dependent hypericin phototoxicity to EMT6 cells have previously shown that 0.5 mM hypericin 1 1.5 J cm22 fluorescent light is non-toxic and that 1.0 mM hypericin 1 1.5 J cm22 fluorescent light produces LD40 toxicity. Intracellular events leading to toxicity were revealed at these doses. Lactate dehydrogenase leakage was elevated for both 0.5 mM and 1.0 mM hypericin 1 light immediately following irradiation. While values eventually returned to control levels for 0.5 mM hypericin 1 light, leakage increased over time for 1.0 mM hypericin indicating reversible and irreversible toxicity, respectively. Increases in lipid and protein oxidation were measured immediately following irradiation; however, these parameters return to control levels within 0.5 h for both doses. Both total cellular ATP levels and cellular respiration were depressed by approximately 50% of control values for 1.0 mM hypericin 1 light. These values were unchanged for 0.5 mM hypericin 1 light. Along with previously reported data demonstrating that light-activated hypericin can inhibit mitochondial succinoxidase in beef heart mitochondria in vitro, these data support oxidative stress-initiated mitochondrial damage as a key target in hypericin phototoxicity. © 1998 Elsevier Science Inc. Keywords—Anticancer agents, Antiviral agents, Hypericin, Photosensitizers, Photodynamic therapy, Oxidative damage, Mitochondria, Free radical
INTRODUCTION
these reports, the cellular mechanism of phototoxicity is not well defined. The mitochondria has been proposed as a potential target of photodynamic action. Hematoporphryin derivative (HPD) photosensitization has been shown to inhibit mitochondrial enzymes involved in oxidative phosphorylation in vitro, including the proton translocating ATPase, cytochrome C oxidase, and succinate dehydrogenase.8,9,10 Later, it was demonstrated that cellular ATP levels decreased in a drug and a light dose-dependent manner with HPD PDT and this decrease was directly related to impairment of mitochondrial function.11 More recently, photoactivated hypericin has been shown to inhibit isolated bovine heart mitochondrial succinoxidase in a drug and a light dose-dependent manner.2 These results were further supported in a study with isolated rat liver mitochondria in which membrane potential and succinoxidase activity were decreased along with respiratory control ratio and ADP/O in the presence of hypericin and light.12
Hypericin is emerging as a good candidate sensitizer for the photodynamic therapy (PDT) of neoplastic diseases. The phototoxic effect of hypericin is thought be oxygendependent, presumably via a type II mechanism involving singlet oxygen.1 Recent data are strongly supportive of this hypothesis,2,3,4 although supporting roles for superoxide anion radical, hydrogen peroxide, and hydroxyl radical have been suggested.4,5 Hypericin generates singlet oxygen at a high quantum yield of 0.746 in comparison with other photosensitizers for which the quantum yield ranges from 0.2– 0.6.7 These reactive oxygen species are assumed to be the agents of hypericin toxicity in EMT6 cells in cell culture since hypericin demonstrated no toxicity under hypoxic or dark conditions.3 Beyond Address correspondence to: Ronald S. Pardini, University of Nevada, Department of Biochemistry MS 330, Reno, NV 89557; Tel: 702-784-4107; Fax: 702-784-1419; E-Mail: [email protected]. 144
Time-course of hypericin phototoxicity
In the present study we establish a time frame to investigate the mode of hypericin phototoxicity for EMT6 mouse mammary carcinoma cells in vitro. The previously reported non-toxic and toxic doses, 0.5 mM and 1.0 mM hypericin, respectively, with 1.5 J cm22 fluorescent light (cell survival: 1.081 6 0.108 and 0.609 6 0.032 respectively)3 were utilized to elucidate the early events of hypericin phototoxicity in EMT6 cells. Lactate dehydrogenase leakage assay was used to measure the level of membrane damage over a 3 h period beginning at the end of the irradiation period. Indications of oxidative stress were measured by lipid peroxidation and protein oxidation assays within this period as well. To further investigate the potential role of mitochondrial impairment in the mechanism of hypericin phototoxicity, we examined the effect of photoactivated hypericin on total cellular ATP concentrations and total cellular respiration in EMT6 cells. METHODS
Chemicals Hypericin was obtained from Atomergic Chemetals (Farmingdale, NY). Powdered Waymouth’s media, neonatal clostrum free calf serum, L-glutamine, trypsin, penicillin-G/streptomycin, dimethylsulfoxide (DMSO), and ATP assay kit were obtained from Sigma Chemical Co. (St. Louis, MO). Dulbecco’s phosphate buffered saline (DPBS) and enzyme grade sodium bicarbonate were obtained from Fisher Scientific (Fair Lawn, NJ). All other chemicals were of analytical grade. EMT6 cell culture The characteristics of the EMT6 mouse mammary carcinoma cell line and techniques used for maintaining cultures have been previously reported.13,14 Cells were grown in flasks as monolayers at 37°C in a saturated humidified environment consisting of 5% CO2 and 95% air. Cells were grown in complete media which is Waymouth’s medium supplemented with 15% neonatal colostrum free calf serum 118 units/ml penicillin G, and 119 mg/ml streptomycin. Tissue culture flasks were seeded with 1 3 105 cells/ml of complete media and all tests were conducted on cells in exponential growth phase. Hypericin exposure Cells were washed thoroughly with DPBS and complete media was replaced with fresh complete media prior to exposing anchored cells to hypericin or to DMSO solvent treatment for 1 h. The final concentration
145
of DMSO in each flask was 0.4% (v/v). At the end of hypericin exposure, cells were washed thoroughly with DPBS. Irradiation All procedures, with the exception of irradiation, were carried out in subdued light (,0.010 mW cm22). Anchored cells in flasks were placed on a 3 mm thick glass 1 cm above 8 Philips cool white 24T12 high output fluorescent bulbs. The light dose was measured with an IL 1700 radiometer (International Light; Newburyport, MA) by integrating the energy signal in mW cm22 over the entire period of irradiation (10.31 6 1.01 min). Over the irradiation period, the light intensity decreased from an initial intensity of 2.74 6 0.17 mW cm22 to a final intensity of 2.14 6 0.21 mW cm22. The light source was extinguished when the total light dose reached 1.5 J cm22. Fluorescent light intensity was similar when measured through plastic flasks used for tissue culture. Temperature measurements recorded in tissue culture flasks during irradiation never exceeded 37°C and decreased to no lower than 31°C by the end of irradiation. Cells in flasks serving as dark controls were placed under a cardboard box at room temperature during the irradiation period. Clonogenic cell survival studies with conditions of irradiation (no hypericin) described here have been previously investigated and demonstrated no substantial alteration in cell survival compared to dark control.3 Therefore, it is assumed that no toxicity occurs in EMT6 mouse mammary carcinoma cells due to irradiation alone. Post-irradiation period incubation Cells either were assayed immediately following the irradiation period (0 h) or were incubated for 0.5 h, 1 h, or 3 h beginning when the complete media was replaced in all flasks. Time between extinguishing the light source and replacing the complete media in all flasks was not more than 5 min. At the end of post-irradiation incubation time, media was decanted off and cells were removed from the flasks manually by scraping. Samples were placed on ice for the remainder of testing. Lactate dehydrogenase (LDH) leakage Following hypericin exposure and rinsing, complete media was replaced and cells were irradiated as described above. Cells were handled as described for the post-irradiation period above except that they remained in the complete media in which they were placed following the hypericin incubation. Immediately following
146
S. A. S. JOHNSON et al.
irradiation, or at the end of the post-irradiation incubation, cells were scraped within the flask and mixed thoroughly with the complete media within the flask. A known volume of media/cell mixture was removed and centrifuged at 1,000 3 g for 5 min. Media supernatant (0.5 ml) was removed to measure the levels of LDH leaked. The balance of the mixture was then sonicated to measure total LDH activity. LDH activity was measured as described by Kornberg.15 Activity of complete media alone was measured and subtracted from the activities of the media supernatant plus the sonicated mixture prior to calculating total percentage of LDH activated leaked. Lipid peroxidation and protein oxidation Following the appropriate incubation time, cells were removed by scraping then were centrifuged at 1,000 3 g for 5 min. Remaining supernatant was decanted and replaced with a lysing buffer (10 mM Tris, 1 mM EDTA, pH 7.5, 2% Triton X-100) then vortexed vigorously and sonicated. Complete preparation of cell extracts took no longer than 20 min from the time of cell removal to sonication of the cell mixture. Thiobarbituric reactive species (TBARS) were measured as an indication of lipid peroxidation as described by Esterbauer and Cheeseman.16 Whole cell homogenates were mixed with 20% (w/v) trichloroacetic acid (half of the original volume) and 0.67% (w/v) thiobarbituric acid (equal amount of the original volume) then heated in a boiling water bath for 10 min. Precipitate was centrifuged out at 10,000 3 g for 10 min. Thiobarbituric reactive species in the supernatant were then measured spectrophotometrically at 535 nm (« 5 1.56 3 105 M21 cm21). Absorbances of lysing buffer controls were subtracted from test sample absorbances. Protein carbonyl groups were measured as an indication of protein oxidation according to the method described by Levine, et al.17 Schiff’s bases were also measured by this procedure. DNA was removed from the fraction by streptomycin precipitation. Remaining proteins were precipitated and the pellet was mixed with 10 mM 3,4-dinitrophenyl hydrazine (DNPH) for 1 h. Excess DNPH was removed with a 1:1 ethanol:ethyl acetate wash mixture and the pellet was re-dissolved in 0.6 M guanidine in 20 mM potassium phosphate buffer, pH 6.3. Solution absorbance was then measured at 390 nm («360 –390 nm 5 22,000 M21 cm21) to determine the presence of protein carbonyl groups. Control sample absorbances without DNPH were subtracted from test absorbances. ATP assay Immediately following irradiation, cells were washed 3 times with DPBS, removed from the flasks with 0.05%
trypsin diluted in DPBS, and then counted with a Coulter counter (Coulter Scientific, Hialeah, FL). Cells were centrifuged at 1,000 3 g for 5 min, and the remaining media supernatant was removed. The cell pellet was then rinsed with DPBS and re-centrifuged at 1,000 3 g for 5 min. DPBS supernatant was removed, and 500 ml of 67 mM Tris, 33.3 mM magnesium sulfate, 6.7 mM EDTA, pH 7.8 was added. The cell mixture was then sonicated and precipitated with 100 ml cold 20% (w/v) trichloroacetic acid. The mixture was neutralized with 6 M potassium hydroxide and centrifuged at 10,000 3 g for 3 min. Total cellular ATP for each test was measured by bioluminescent assay (Sigma, FL-AA) using Na2ATP in a standard curve on a Lab System Luminoskan Luminometer. Test sample ATP concentrations were compared to ATP levels in control cells receiving no treatment during the same isolation to get percent of control values. Cellular respiration Immediately following irradiation, cells were washed 3 times with DPBS, removed from the flasks with 0.05% trypsin diluted in DPBS, and then counted with a Coulter counter (Coulter Scientific, Hialeah FL). Cells were centrifuged at 1,000 3 g for 5 min and remaining media supernatant was removed. Cells were diluted to a final concentration of 2.5 3 106 cells/ml in complete media. Complete Waymouth’s media contains a final concentration of 5 g/l glucose and essential amino acids. Other than this, no other substrates of glycolysis, electron transport, or oxidative phosphorylation were supplemented during the measurement of oxygen consumption. Cells were kept on ice until oxygen consumption was measured using a Gilson water jacketed reaction chamber (Middleton, WI) set at 37°C and equipped with a Yellow Springs Instruments model 53 oxygen monitor (Yellow Springs, OH) and a Clark electrode. Complete media oxygen consumption was measured prior to addition of cells and was not significant. Cellular oxygen consumption was measured for a minimum of 5 min. Protein assays Protein concentrations were determined using the BCA kit (Pierce, Rockford, IL). Statistical analysis Data were analyzed by analysis of variance using SAS Systems for Windows release 6.11 TS Level 0040 computer software (SAS Institute Inc., Cary, NC). If the analysis exhibited a significant difference, data were
Time-course of hypericin phototoxicity
147
Fig. 1. Hypericin toxicity towards EMT6 mouse mammary carcinoma cells. Percent of LDH leaked measured immediately following irradiation with 1.5 J cm22 fluorescent light. Cells were incubated with hypericin 1 h prior to irradiation. Statistical differences based on Tukey–Kramer least square means comparisons: alight vs. dark control; bhypericin vs. DMSO control.
further analyzed by Tukey–Kramer least square means comparisons to discern significant differences among the means ( p , .05). RESULTS AND DISCUSSION
Phototoxicity to EMT6 mouse mammary carcinoma cells has been established.3 The survival curve generated in that study showed that tolerance for hypericin phototoxicity extended up to 0.5 mM hypericin with light and then diminished sharply thereafter. LDH leakage was measured in EMT6 cells immediately following irradiation to determine the amount of membrane damage that occurred during the photoactivation period (Fig. 1). The effect was dose and light-dependent with the highest dose measured, 2.5 mM hypericin 1 light, sustaining enough membrane damage to demonstrate 91.4% of total activity leaked indicating almost instantaneous cytotoxicity for these cells. The levels of leakage experienced by cells treated with 0.5 mM hypericin 1 light and 1.0 mM 1 light immediately following irradiation were elevated over their dark controls but were not statistically different from one another (Fig. 1 and 2, p , 0.99). We know from the previous clonogenic assays that 0.5 mM
hypericin 1 light is not toxic to EMT6 cells while 1.0 mM hypericin 1 light is marginally toxic.3 This implies that photoactivated hypericin elicited enough damage at the non-toxic dose to allow elevated membrane leakage but that the dose was not strong enough to evoke eventual toxicity while the higher dose, experiencing similar levels of initial membrane damage, was unable to recover. Given this information, we decided that these two doses could be targeted to dissociate key events leading to eventual hypericin phototoxicity in EMT6 mouse mammary carcinoma cells. The equivalent levels of initial LDH leakage between the non-toxic and marginally toxic doses led us to focus on later events in the time-course of eventual hypericin toxicity. Following the irradiation period, cells were incubated for 0.5, 1, and 3 h. Consistent with the clonogenic data, total activity leakage increased over a time period of 3 h for cells exposed to 1.0 mM hypericin 1 light (Fig. 2). In contrast, cells exposed to 0.5 mM hypericin 1 light demonstrated no further indication of membrane damage and intracellular LDH levels were restored back to the level of the controls (Fig. 2). The effect of phototoxicity with hypericin is oxygendependent.3 The fact that hypericin is known to produce
148
S. A. S. JOHNSON et al.
Fig. 2. LDH leakage following treatment with hypericin 6 1.5 J cm22 fluorescent light from EMT6 mouse mammary carcinoma cells. Cells were prepared for assay immediately following irradiation (0 h) or allowed to incubate in complete media following the irradiation period (0.5 h, 1 h, or 3 h). Values are means 6 SEM of 5–35 independent determinations. Statistical differences based on Tukey–Kramer least square means comparisons: alight vs. dark control; bhypericin vs. DMSO control; cadditional relevant statistical differences: 1.0 mM hypericin 1 light: 0 h vs. 1.0 mM hypericin 1 light: 0.5 h, 1 h, 3 h.
high quantum yields of singlet oxygen6 along with superoxide anion radical4,5 suggests a mechanism of phototoxicity in which the generation of reactive oxygen species causes oxidative damage which must be demonstrated cellularly. Evidence of oxidative damage was investigated by measuring indices of lipid peroxidation and protein oxidation on the same time-course basis used previously. Both TBARS and protein carbonyls were elevated immediately following irradiation for both 1.0 mM and 0.5 mM hypericin 1 light but returned to control levels within 0.5 h (Fig. 3). While lipid peroxidation levels were higher for 1.0 mM hypericin 1 light than for 0.5 mM hypericin 1 light, protein oxidation was equivalent for both doses. These results are supportive of previous studies which have demonstrated hypericin’s ability to induce lipid peroxidation18,19 and further establishes that oxidative damage occurs at a cellular level as well. The fact that these general measures of oxidative damage are absent after 0.5 h of the post-irradiation period suggests that EMT6 cells were able to repair this
damage for both doses. Yet despite this, 1.0 mM hypericin 1 light is toxic to EMT6 cells and membrane damage is evident (Fig. 1, 2). Hypericin toxicity has been consistently shown to involve reactive oxygen species as antioxidants and free radical scavengers, in addition to hypoxic conditions, have all been shown to protect cells from photoactivated hypericin.3,20 This implies that reactive oxygen species generated by photoactivated hypericin may have more discrete targets which are unable to be repaired following photo-oxidative damage at 1.0 mM hypericin 1 light in EMT6 cells. It is also possible that damage to critical targets arises from the initial oxidative damage. Mitochondria are considered a primary target of phototherapy,21,22 and hypericin association with cellular membranes has been observed.3,23 As seen in Figure 4A, total cellular ATP levels in EMT6 mouse mammary carcinoma cells were decreased to 53.6 6 2.4% of controls by 1.0 mM hypericin 1 light. This decrease is light-dependent and is consistent with previous cytotox-
Time-course of hypericin phototoxicity
Fig. 3. Lipid peroxidation (panel A) and protein oxidation (panel B) following treatment with hypericin 6 1.5 J cm22 fluorescent light in EMT6 mouse mammary carcinoma cells. Lipid peroxidation was measured as thiobarbituric reactive species absorbing at 532 nm. Protein oxidation was measured as protein carbonyls reacting with 2,4-dinitrophenyl hydrazine. Cells were prepared for assay immediately following irradiation (0 h) or allowed to incubate in complete media following the irradiation period (0.5 h, 1 h, or 3 h). Values are means 6 SEM of 3–9 independent determinations. Statistical differences based on Tukey–Kramer least square means comparisons: alight vs. dark control; bhypericin vs. DMSO control; cadditional relevant statistical differences: Lipid peroxidation: 1.0 mM hypericin 1 light: 0 h vs. 1.0 mM hypericin 1 light: 0.5 h, 1 h, 3 h; 0.5 mM hypericin 1 light: 0 h vs. 0.5 mM hypericin 1 light: 0.5 h, 1 h, 3 h. Protein oxidation: 1.0 mM hypericin 1 light: 0 h vs. 1.0 mM hypericin 1 light: 0.5 h, 1 h, 3 h; 0.5 mM hypericin 1 light: 0 h vs. 0.5 mM hypercin 1 light: 0.5 h, 1 h, 3 h; DMSO 1 light: 1 h vs. DMSO 1 light: 3 h.
149
150
S. A. S. JOHNSON et al.
Fig. 4. Total cellular ATP (panel A) and total cellular respiration (panel B) in EMT6 mouse mammary carcinoma cells following treatment with hypericin 6 1.5 J cm22 fluorescent light. Cells were incubated with hypericin 1 h prior to irradiation. ATP data is present as percent control (means 6 SEM, n 5 5–14) where tested ATP levels were compared to total ATP extracted from control cells (no treatment) during the same extraction. For cellular respiration, cells were suspended in complete Waymouth’s media containing a final concentration of 5 g/l glucose and essential amino acids. Other than this, no other substrates of glycolysis, electron transport, or oxidative phosphorylation were added during the measurement of oxygen consumption. Cellular respiration (means 6 STD, n 5 5–9) was measured as oxygen consumption with a Gilson water jacketed reaction chamber equipped with a Yellow Springs Instrument model 53 oxygen monitor and a Clark electrode. Statistical differences based on Tukey–Kramer least square means comparisons, alight vs. dark control; bhypericin vs. DMSO control.
icity data. In addition, 1.0 mM hypericin 1 light proportionally inhibited total cellular respiration (Fig. 4B) to about half of control respiration rates. Cells treated with 1.0 mM hypericin consumed only 0.00076 6 0.00001 matoms O/min per 106 cells while cells treated with DMSO 1 light consumed 0.00143 6 0.00013 matoms
O/min per 106 cells. In contrast, non-toxic 0.5 mM hypericin with and without light had no effect on total cellular ATP levels and total cellular respiration (Fig. 4A, B). These observations are supportive of previous data which demonstrated hypericin’s ability to impair mitochondrial function in vitro in a drug and light-dose
Time-course of hypericin phototoxicity
dependent manner.2,12 More recently, release of mitochondrial hexokinase activity from the mitochondria as well as inhibition of this activity has been induced by photoactivated hypericin in human glioma cells.24 This effect was accompanied by a hypericin-induced decrease in intracellular pH in addition to decreases in ATP and glucose-6-phosphate. While the hexokinase release from mitochondria was attributed to hypericin-dependent intracellular acidification, hexokinase inhibition was oxygen-dependent.24,25 The dual effects of oxidative damage and change in intracellular pH elicited by photoactivated hypericin could be the precise combination which makes this drug, “one of the most powerful photodynamic agents in nature.”26 In combination with the current observations, it is reasonable to extend the possible role of mitochondria as a target of oxygen-dependent hypericin phototoxicity from a sub-cellular to a cellular level. CONCLUSIONS
Photoactivated hypericin initiates a series of oxidative stress events at a cellular level. At the non-toxic and toxic doses tested, two indicators of oxidative stress, lipid peroxidation and protein oxidation, were initially elevated then returned to control levels for both doses in EMT6 cells. Membrane damage is evident initially for nontoxic 0.5 mM hypericin 1 light and toxic 1.0 mM hypericin 1 light. Cells exposed to 0.5 mM hypericin 1 light repaired damage within 3 h, consistent with clonogenic toxicity data. Membrane damage for 1.0 mM hypericin 1 light persisted for the duration of the experiment. These doses will be useful in differentiating the key events that ultimately end in cellular toxicity. One possible discrete target of hypericin phototoxicity may be mitochondria. At 1.0 mM hypericin 1 light, cellular respiration was depressed by 47% compared to DMSO 1 light controls. Similarly, cellular ATP concentrations were depressed by 47% compared to untreated controls by the same hypericin treatment. These findings suggest that photoactivated hypericin interact at the mitochondrial level in EMT6 cells. This conclusion is supported by our recent report that MnSOD was initially elevated and subsequently depressed over a 3 h time period following treatment with 1.0 mM hypericin 1 light.27 Further investigation of photoactivated hypericin’s effect on glycolysis, electron transport, and oxidative phosphorylation in whole cells will help further elucidate the role of mitochondria in hypericin’s mechanism of cellular toxicity. Acknowledgements—Funding for these studies was provided in part by The Grand Chapter of Nevada, Order of the Eastern Star and the Howard Hughes Medical Institute Undergraduate Biological Sciences Education Program, Grant #71192-516901. We are grateful to Laurent
151
Miccoli for helpful discussion. We thank Megan Probst for technical assistance and Dr. Chris Pritsos for providing EMT6 cells for our culture experiments.
REFERENCES 1. Duran, N.; Song, P.-S. Yearly Review: Hypericin and its photodynamic action. Photochem. Photobiol. 43:677– 680; 1986. 2. Thomas, C.; MacGill, R. S.; Miller, G. C.; Pardini, R. S. Photoactivation of hypericin generates singlet oxygen in mitochondria and inhibits succinoxidase. Photochem. Photobiol. 55:47–53; 1992. 3. Thomas, C,; Pardini, R. S. Oxygen dependence of hypericin induced phototoxicity to EMT6 mouse mammary carcinoma cells. Photochem. Photobiol. 55:831– 837; 1992. 4. Diwu, Z.; Lown, J. W. Photosensitization with anticancer agents 17. EPR studies with photodynamic action of hypericin: formation of semiquinone radical and activated oxygen species on illumination. Free Radical Biol. Med. 14:209 –215; 1993. 5. Hadjur, C.; Jeunet, A.; Jardon, P. Photosensitization by hypericin: electron spin resonance (ESR) evidence for the formation of singlet oxygen and superoxide anion radicals in an in vitro model. J. Photochem. Photobiol. B.: Biol. 26:67–74; 1994. 6. Jardon, P.; Lazortchak, N.; Gautron R. Formation d’oxygene singluet 1Dg photosensibilisee par l’hypericine. Caracterisation et etude du mecanisme par spectroscopie laser. J. Chim. Phys. 84: 1143–1145; 1987. 7. Pass, H. I. Photodynamic therapy in oncology: mechanism and clinical use. J. Natl. Cancer Inst. 85:443– 456; 1993. 8. Gibson, S. L.; Hilf, R. Photosensitization of mitochondrial cytochrome C oxidase by hematoporphyrin derivative and related porphyrins in vitro and in vivo. Cancer Res. 43:4191– 4197; 1983. 9. Hilf, R.; Small, D. B.; Murant, R.; Leaky, P. B.; Gibson, S. Hematoporphyrin derivative-induced photosensitivity of mitochondrial succinate dehydrogenase and selected cytosolic enzymes of R3230AC mammary adenocarcinomas of rats. Cancer Res. 44:1483–1488; 1984. 10. Perlin, D. S.; Murant, R. S.; Gibson, S. L.; Hilf, R. Effect of photosensitization by hematoporphyrin derivative on mitochondrial adenosine triphosphatase-mediated proton transport and membrane integrity of R3230AC mammary adenocarcinoma. Cancer Res. 45:653– 658; 1985. 11. Hilf, R.; Murant, R. S.; Narayan, U.; Gibson, S. L. Relationship of mitochondrial function and cellular adenosine triphosphate levels to hematoporphyrin derivative-induced photosensitization in R3230AC mammary tumors. Cancer Res. 46:211–217; 1986. 12. Utsumi, T.; Okuma, M.; Kanno, T.; Takehara, Y.; Yoshioka, T.; Fujita, Y,; Horton, A. A.; Utsumi, K. Effect of the antiretroviral agent hypericin on rat liver mitochondria. Biochem. Pharmacol. 50:655– 662; 1995. 13. Rockwell, S. C.; Kallman, R. F.; Fajardo, L. F. Characteristics of a serially transplanted mouse mammary tumor and its tissueculture-adapted derivative. J. Natl. Cancer Inst. 49:735–747; 1972. 14. Rockwell, S. C. In vivo-in vitro tumor systems: new models for studying the response of tumors to therapy. Lab. Anim. Sci. 27: 831– 851; 1977. 15. Kornberg, A. Lactate dehydrogenase of muscle. Methods Enzymol. 1:441– 443; 1955. 16. Esterbauer, H.; Cheeseman, K. H. Determination of aldehydic lipid peroxidation products: malondialdehyde and 4-hydroxynonenal. Methods Enzymol. 186:407– 421; 1990. 17. Levine, R. L.; Garland, D.; Oliver, C. N.; Amici, A.; Climent, I.; Lenz, A.-G.; Ahn, B.; Shaltiel, S.; Stadtman, E. R. Determination of carbonyl content in oxidatively modified proteins. Methods Enzymol. 186:464 – 478; 1990. 18. Hadjur, C.; Richard, M.-J.; Parat, M.-O.; Jardon, P.; Favier, A. Photodynamic effects of hypericin on lipid peroxidation and antioxidant status in melanoma cells. Photochem. Photobiol. 64:375– 381; 1996. 19. Knox, J. P.; Dodge, A. D. Isolation and activity of the photodynamic pigment hypericin. Plant Cell Environ. 8:19 –25; 1985.
152
S. A. S. JOHNSON et al.
20. Hadjur, C.; Richard, M. J.; Parat, M. O.; Favier, A.; Jardon, P. Photodynamically induced cytotoxicity of hypericin on human fibroblast cell line MRC5. J. Photochem. Photobiol. B: Biol. 27: 139 –146; 1995. 21. Hilf, R.; Gibson, S. L.; Penney, D. P.; Checkler, T. L.; Bryant, R. C.; Mitochondria in cancer cells as targets of photodynamic therapy. Proc. Soc. Photo-Opt Instrum. Eng. 847:2–10; 1989. 22. Salet, C.; Moreno, G. New trends in photobiology (Invited Review): photosensitization of mitochondria. Molecular and cellular aspects. J. Photochem. Photobiol. B: Biol. 5:133–150; 1990. 23. Miskovsky, P.; Sureau, F.; Chinsky, L.; Turpin, P.-Y. Subcellular distribution of hypericin in human cancer cells. Photochem. Photobiol. 62:546 –549; 1995. 24. Miccoli, L.; Sureau, F.; Dutrillaux, B.; Poupon, M.-F. Hypericin disrupts the hexokinase mitochondria-binding and affects the pHdependent metabolism of human gliomas. Proc. Am. Assoc. Cancer Res. 38:598; 1997. 25. Miccoli, L.; Oudard, S.; Sureau, F.; Dutrillaux, B.; Poupon, M.-F. Intracellular pH governs the subcellular distribution of hexokinase in a glioma cell line. Biochem. J. 313:957–962; 1996.
26. Walker, E. B.; Lee, T. Y.; Song, P.-S. Spectroscopic characterization of the stentor photoreceptor. Biochim. Biophys. Acta 587:129 – 144; 1979. 27. Johnson, S. A. S.; Pardini, R. S. Antioxidant enzyme response to hypericin in EMT6 mouse mammary carcinoma cells. Free Radical Biol. Med. 24:817– 826; 1998.
ABBREVIATIONS
DNPH— 4-dinitrophenyl hydrazine DMSO— dimethyl sulfoxide DPBS—Dulbecco’s phosphate buffered saline HPD— hematoporphyrin derivative LDH—lactate dehydrogenase PDT—photodynamic therapy TBARS—thiobarbituric reactive species