(CANCER RESEARCH 53. 2571-25KO. June I. I993|

Phototoxicity, Redox Behavior, and Pharmacokinetics of BenzophenoxazineAnalogues in EMT-6 Murine Sarcoma Cells1

Louis Cincotta,2 James W. Foley, and Anthony H. Cincotta

Rowland Inaiiate for Science. Cambridge 02142 ¡L.C., J. W. F.¡,and Wellman Laboratories of Pholomedicine. Massachusetts General Hospital. Harvard Medical School.Department of Dermatology. Boston 02114 ¡A.H. C.]. Massachusetts

ABSTRACT

Structural modifications to the photoinactive benzophenoxazine Nileblue A have led to three novel derivatives which include 5-ethylamino-9-diethylaminobenzo[alphenoxazinium (EtNBA), 5-ethylamino-9-diethyl-aminobenzo[a]phenothiazinium (EtNBS), and 5-ethylamino-9-diethyl-

aminobenzo[a]phcnosclenazinium (EtNBSe) chlorides. The incorporationof sulfur and selenium into the benzophenoxazine moiety results in lipo-philic, red-absorbing (650-660 nm) chromophorcs which possess signifi

cantly increased singlet oxygen yields (0.025 and 0.65, respectively, compared to 0.005 for EtNBA). This study examines the photosensitizingefficacies and pharmacokinetics in vitro in the EMT-6 murine mammary

sarcoma cell line as well as the physicochemical, photochemical, and redoxproperties of these new analogues. Comparisons with Photofrin II, theonly photosensitizer available clinically, were made in an attempt to highlight their different pharmacological characteristics. The photodynamicactivity of the benzophenoxazine dyes correlates with their ability to generate the phototoxin singlet oxygen and increases in the following order:EtNBA < EtNBS « EtNBSe. At an extracellular dye concentration of 0.5

UM,the light dose required to kill approximately 50% of the cells was 2.0and <0.5 J/cm2 for the sulfur and selenium dyes, respectively. The light

dose required to kill approximately 50% of the cells for both EtNBA andPhotofrin II could not be determined because of their weak phototoxiceffect under these conditions. At a light dose of 3.3 J/cm2, EtNBSe is

approximately 1000 times more phototoxic than Photofrin II. All threebenzophenoxazine derivatives are characterized by a similar uptake/effluxpattern in vitro consisting of a rapid and extensive cellular accumulationfollowed by a slow efflux rate. Contrary to their rapid uptake, 50% of theaccumulated EtNBS and EtNBSe is retained intraccllularly after a 6-hperiod in dye-free medium. Video-enhanced fluorescence microscopy cor

roborates the rapid uptake measurements as well as indicating the intra-

cellular localization of the dyes in both living and thermally inactivatedcells. Low extracellular dye concentrations (0.05 IIMiresult in a punctatefluorescence pattern in the perinuclear region, while higher dye concentrations i O.I UM) lead to additional fluorescence in the cytoplasm, cy-

tomembranes, and other organdíes but apparently not the nucleus. Absorption spectrometry revealed that living cells rapidly reduce the dyes totheir colorless leuko form (photoinactive) if oxygen is not readily availablein the environment. It is shown that the cellular reduction is an enzymaticprocess and that an oxygen-free and cell-free medium containing both the

coenzyme NADH and the hydride transfer enzyme diaphorasc is capableof reducing the dyes to the colorless leuko form. EtNBSe exhibited thefastest rate of reduction under these conditions. In addition, the benzophenoxazine derivatives also undergo a light-induced reduction under anaer

obic conditions in the presence of NADH; the rates increase in the following order: EtNBA < EtNBS « EtNBSe. Rcoxidation of the leuko formoccurs upon introduction of oxygen. This investigation demonstrates thatthe benzophenoxazine chalcogcn analogues are a unique class of photo-

chemotherapeutic agents.

Received 12/15/92; accepted 3/26/93.The costs of publication of this article were defrayed in part by the payment of page

charges. This article must therefore be hereby marked advertisement in accordance with18 U.S.C. Section 1734 solely to indicate this fact.

1Funded by the Rowland Institute for Science and the Wellman Laboratories for

Photomedicine. Massachusetts General Hospital.2 To whom requests for reprints should be addressed, at the Rowland Institute for

Science. 100 F.dwin Land Boulevard. Cambridge. MA 02142.

INTRODUCTION

PDT,3 an experimental modality for treating solid tumors (1), is

currently undergoing clinical evaluation. The therapy involves thesystemic administration of tumor-localizing photosensitizers that can

inactivate malignant tissue when irradiated with light of the appropriate wavelength. The most widely studied PDT drugs, HPD and PII (apurified fraction of HPD), suffer from several limitations: (a) bothexhibit a low absorption coefficient in the region where light penetrates tissue most efficiently (600-900 nm); (b) both are complex

mixtures of porphyrin ether and ester oligomers; and (c) the prolongedretention of these photosensitizers in the skin leads to dermal photo-sensitization that can persist for months. Mechanistically, these por-

phyrins appear to cause tumor damage primarily via the acute destruction of the microvasculaturc in the tumor and not by the directphotodestruction of the tumor cells. This results in a rapid shift ofsome cells into hypoxia where they are potentially protected fromfurther PDT damage as a consequence of the oxygen depletion. Nutritional resupply to these still viable tumor cells through diffusion orangiogenesis can rapidly repopulate the tumor (for a review see Ref.2).

The recognized limitations of HPD as a photosensitizer for photo-

dynamic therapy have stimulated the search for new, more efficaciousphototoxic drugs (for reviews see Rcfs. 2-5). Surprisingly, there have

been few systematic attempts to develop these drugs from classes ofdyes other than the extended porphyrin family (i.e., chlorins, purpurins, phthalocyanines, etc.). By examining photosensitizers with inherently different physicochemical and pharmacological properties wehope to formulate PDT agents having a fundamentally different modeof action when compared to the porphyrins. To this end, we haveconcentrated our efforts on the development of novel, cationic, pho-

tochemotherapeutic drugs belonging to the benzophenoxazine familyof chromophores (6-11). Our interest in this family of dyes was

stimulated by a series of early investigations initiated by Lewis el al.(12, 13) which demonstrated the propensity of benzophenoxa/ines toselectively stain tumors. Unfortunately, the commercially availablematerials studied by Lewis were found to be inefficient photosensitizers both in vitro and in vivo (6).

We attribute the low level of phototoxicity of benzophenoxazines tothe documented inability of this class of dyes to efficiently forni tripletstates during irradiation (14). It is the long life of the triplet state (10~6to 10~3 s) which allows sufficient time for the transfer of energy to

occur between the excited photosensitizer and ground state oxygenthat is responsible for the generation of the cytotoxin 'O2 (type II

reaction), a molecule that is believed to be responsible for most PDTeffects (15). Consequently, any approach for the design of more efficient benzophenoxazine photosensitizers must include a mechanismfor enhancing their triplet state quantum yields. Chalcogens (i.e., S,Se, Te) are commonly incorporated into molecules having low intrinsic rates of intersystem crossing between the excited singlet and triplet

'The abbreviations used are: PDT. photodynamic therapy; ED.â„¢,dose required to kill

50% of cells; EtNBA. 5-ethylamino-9-dieihylamino-ben7o[u]phenoxazinium chloride;EtNBS. 5-ethylamino-9-diethylaminobenzo[i;]phenothiazinium chloride; ElNBSe, 5-eth-ylamino-9-diethylaminobenzo[u]phenoselenazinium chloride; HBSS. Hanks' balanced

salt solution; HPD. hematoporphyrin derivative; PII. Photofrin II.

2571

Research. on August 20, 2021. © 1993 American Association for Cancercancerres.aacrjournals.org Downloaded from

PHOTOTOXICITY OF BEiNZOPHENOXAZINE ANALOGUES

state manifolds in order to enhance triplet yields via their large spin-orbital coupling constants ("heavy-atom effect") (14, 16). We have

exploited this strategy with the benzophenoxazines and replaced theoxygen atom in EtNBA chloride with sulfur and selenium to give theanalogues EtNBS and EtNBSe chloride, respectively. The presentstudy investigates both the pharmacokinetics and phototoxic effects ofthese novel derivatives in EMT-6 cells and compares them with PII.

Our observation that pelletized, live cells containing these new dyeswere colored blue and that they slowly decolorized with the exceptionof those at the air/cell interface prompted us to investigate both theenzymatic and photoredox behaviors of the chromophores. A perusalof the pertinent literature (17-22) supported the assumption that the

reduction of the new dyes to a colorless leuko form was occurringunder hypoxic conditions and that the reaction was reversible in thepresence of sufficient oxygen. If this were true it would have important ramifications not only for the evaluation but also for the efficacyof the photosensitizers both in vilro and in vivo. Therefore, severalexperiments were designed in order to establish whether the dyes werebeing reduced by live EMT-6 cells, whether the dyes could undergo a

photoreduction, whether both types of reduction were reversible in thepresence of oxygen, and if the ubiquitous cellular coenzyme/reductantNADH might be involved in the process.

MATERIALS AND METHODS

Chemicals. EtNBA and ElNBS were synthesized according to proceduresdescribed in U.S. patent 4.962.197. General synthetic schemes for EtNBSe(23) as well as for the other dyes are well documented (24. 25). All dyes werepurified by medium pressure (100 psi) liquid chromatography. Silica gel(Woelm 32-63) was used as the solid phase, eluting with a linear gradient of

méthylènechloride:methanol (100:0 to 90:10). They were homogeneous bythin-layer chromatography and high-field nuclear magnetic resonance spec-

troscopy (JEOL 400 MHz). Photofrin II was kindly provided by Quadra LogicTechnologies, Inc. (Vancouver. BC, Canada). Nile red was purchased fromAldrich Chemical Company (Milwaukee, WI) and used without further purification. Meldola blue was obtained from Pfaltz and Bauer (Waterbury, CT)and purified by the above procedure prior to use.

Cell Culture. The EMT-6 tumor cell line (a murine mammary sarcoma)

was obtained from Dr. E. Lord of the University of Rochester Cancer Center(Rochester. NY) and maintained according to the protocol of Rockwell et al.(26). Radiobiological and immunological characteristics of this cell line havebeen described previously (27). The cells were cultured in RPMI 1640 (SigmaChemical Co.. St. Louis. MO) supplemented with 10% fetal bovine serum(Sigma). 10* units/liter penicillin, 100 mg/liter streptomycin, and 25 mM sodium bicarbonate. The cells were incubated at 37°Cin a humidified 95%

air:5% CO2 atmosphere.Dye Uptake Studies. Subconfluent cell cultures in log phase of growth

were utilized for dye uptake studies at approximately 3 x IO6 cells/25 cm2 orapproximately 1 x 10* cells/9.6 cm2. Medium was aspirated from cultures and

replaced with 2 ml of phenol red-free HBSS containing 3 \IMdye. Cells wereincubated at different time intervals (5, 10, 20, and 40 min) in the dark at 37°C,

and the cellular dye concentration was determined at these intervals. To examine the effect of dye concentration on uptake, cells were incubated at 37°C

for 30 min with 2 ml of HBSS containing 6 different extracellular dye concentrations: 0.5, 1.0, 5.0, 10, 50. and 100 \IM.The uptake of PII was examinedonly at one time point (30 min incubation) and concentration (1.0 (IM)sincethese were the most stringent conditions used during the cell killing experiments. For the sake of comparisons it should be noted that 1.0 uM of dyecorresponds to approximately 0.4 and 0.6 pg/ml for the benzophenoxazines andPH. respectively. Since the formula weight of Photofrin II is uncontrolled, thevalue of the monomer (600) was used in these experiments.

Dye Retention Studies. The retention characteristics of the new chromophores and PII were determined by utilizing subconfluent cell cultures inlog phase of growth (approximately 3 X IO6 cells/25 cm2). The cells wereincubated with 2 ml of HBSS containing 3 (IMdye for 30 min at 37°Cto allow

uptake of the dye. The medium was removed, and the culture flasks werewashed twice and overlayered with RPMI 1640. Cellular dye concentrations

were determined at various time intervals for up to 6 h. To reduce the uptake

of extracellular dye. the RPMI 1640 was replaced every hour throughout thewashout period.

Cellular Dye Extraction and Quantitation. Cells grown in culture flaskswere washed twice with HBSS. trypsinized. placed in Eppendorf tubes, andpelletized by centrifugation (8000 x g). Benzophenoxazines were extractedfrom the resultant pellet with 1 ml of solvent (1:1 methanol:chloroform containing 5 pi acetic acid/ml). The amount of dye present was then quantitatedspectrophotometrically (Perkin-Elmer Lambda 5 UV/VIS). Extraction studieshave shown that 40-50% of the administered dye is actually taken up by the

cells, while the remainder can be accounted for by either dye bound to the

plastic surfaces and/or by dye remaining in the extracellular media. PII wasextracted from the cells in 0.1 N NaOH, and fluorescence was measured in 0.1N NaOH at 625 nm with excitation at 400 nm using a Fluorolog 2 spectroflu-

orometer (Spex Industries, Edison, NJ). A standard concentration curve wasproduced by adding known amounts of PII to O.I N NaOH containing a knownamount of cells. Prior to centrifugation an aliquot was taken for the determination of total protein by a modified Lowry method using a commerciallyavailable kit (Sigma no. 690 A). All manipulations of cells exposed to photo-

sensitizer were performed in subdued light.Sensitization and Irradiation. The photocytotoxicity of the various ben-

zophenoxazine analogues as well as PII were evaluated by varying both thelight dose and the dye concentration. An appropriate number of cells (0.5-2.0x I04/0.5 ml) were seeded in a 24-well, flat bottomed. Linbro plate. Forty-

eight h later, subconfluenl monolayers were incubated for 30 min in the darkwith 1 ml HBSS containing either 1.0, 0.5, 0.1, 0.05. or 0.01 UMsensitizer orwith HBSS alone at 37°C.Following incubation, all cultures within each group

(dye or HBSS) were washed twice with HBSS and either subjected to irradiation (10 min; 3.3 ¡/cm2:590-700 nm) or maintained in the dark in HBSS at20°C.The effect of light dose on phototoxicity was determined similarly. Cells

were treated with 0.5 PM dye in 1 ml of HBSS as above and irradiated with0.66. 1.32. 1.98. 2.64. or 3.3 J/cm2. After treatment all cultures were placedback into growth media at 37°Cuntil they were assayed for cell viability. The

light source was a Polaroid 610 slide projector and is described in detailelsewhere (6).

Cell Viability. Cell viability was determined 48 h following irradiation via['H]thymidine incorporation (ICN Radiochemicals, Irvine. CA) into cellularDNA. Cells were exposed to ['H]thymidine (2 pCi/0.5 ml; specific activity, 50Ci/mw) in supplemented minimum Eagle's medium for 6 h, then washed three

times with HBSS and three times with 5% trichloroacetic acid to precipitate theDNA. Cellular precipitates were dissolved in 0.5 N sodium hydroxide andcounted for 'H content. The percentage of viable cells following irradiation

was calculated using the formula:

(dpm of cells + dye + light)

(dpm of cells + dye + dark)x 100

where dpm is disintegrations per minute. Similarly, the dark toxicity of thedyes was calculated as follows:

(dpm of cells + dye + dark)

(dpm of cells + dark)x 100

Statistical differences in percentage of cell viability between groups weredetermined by unpaired Student's / test. Freshney (28) points out that *H

nucleotide incorporation is a reliable indicator of viability if an appropriaterecovery period (several cell population doubling times) is allowed followingPDT prior to incorporation measurements. The EMT-6 cell line has a doublingtime of —¿�12h. A study (10) of benzophenoxazine cytotoxicity using thestandard clonogenic assay was found to be consistent with our ['Hjthymidine

uptake results. Also, Richter et al. (29) have demonstrated an equivalencybetween the ['Hjthymidine uptake and the standard clonogenic assay for

viability during the phototoxic evaluation of benzoporphyrin derivatives. Priorto the quantitative evaluation of phototoxic effects the cells were observed withlight microscopy. Typically, cell cultures that had sustained serious photochemical damage were found to have severely depleted populations; most remainingcells presented distorted morphologies including cytoplasmic blebbing andvacuolization.

Cellular Reduction in Dye. Subconfluent cell cultures in the log phase ofgrowth (approximately 3 X IO6 cells/25 cm2) were treated with 2 ml HBSS

2572

Research. on August 20, 2021. © 1993 American Association for Cancercancerres.aacrjournals.org Downloaded from

PHOTOTOXICITY OF BKNZOPHF.NOXAZINE ANALOGUES

containing 3 UMdye at 37°Cfor 30 min in the dark, washed with HBSS.

trypsinized. and placed in clear 1.5-ml microcentrifuge tubes. The cells were

pelleti/.ed by centrifugaron (10.000 X g) for 2 min. the HBSS and trypsin were

removed by aspiration, and the cells were layered with 0.5 ml of silicone oil inorder to exclude air. The pellctized cells were placed in a special holder (Fig.

1) coupled by fiber optics to a reflective absorption spectrometer (HewlettPackard 8452-A) fitted with a photodiode array detector. Recordings of the

absorption spectrum were made every 5 min. A spectrum of cells containing no

dye was obtained as a control for endogenous cellular absorption as well asspectral scattering effects. The final spectra were obtained by subtracting this

spectrum from the cells plus dye spectra.Reduction of Dye in Solution. A cell-free system was devised that was

similar to the one of Bennetto and Stirling (17) in order to evaluate theenzymatic reduction of the new chromophores in the absence of light and

oxygen. Because of the ubiquitous nature and importance of the coenzymeNADH as a reductant in cellular processes, it was chosen as the electron donor.The enzyme NADH diaphorase (type II-L from CI. kluyvcrï,Sigma) acts as a

hydride transfer agent between the reductant NADH and the benzophenoxazine

electron acceptor. To a cuvet (10 mm light path) equipped with a septum capwas added 2.0 ml of dye (10 UMin HBSS) and 0.2 ml of NADH solution (1

mg/ml HBSS). The solution was deaerated by purging with argon gas in thedark for 15 min. After the injection of 0.2 ml of argon-purged diaphorasesolution (I mg/ml in HBSS) the disappearance of dye was monitored spectro-

photometrically at room temperature over regular time intervals. The spectro-

photometer was equipped with shutters and a magnetic stirrer to ensure com

plete darkness and homogeneity of solutions between measurements. Theindependent effect of cither NADH or diaphorase with dye was also measured

under the same conditions.

The photoreduction of the chromophores was also examined using NADHas the reductant. To a cuvet equipped with a septum cap was added 2.0 ml of

dye (10 UMin HBSS) and 0.3 ml of NADH solution (1 mg/12 ml HBSS). Thesolution was purged with argon gas in the dark for 15 min. Absorption readingsat the dye's maximum were taken every minute for 5 min in the dark, and then

the solution was exposed to red light (590-700 nm. 0.048 W/cm2) and absorp

tion readings taken every 30 s (5 s for EtNBSe) for an additional 5 min.Microscopy. The cellular localization of the new dyes was determined by

the use of conventional and fluorescence microscopies. The imaging systemconsisted of a Nikon diapot inverted microscope equipped for brightfield.

phase contrast, and epifluorescence illumination using a xenon arc sourcecombined with a Hamamatsu silicon intensified target video camera and Argus

10 image processor (Hamamatsu Photonics K. K.. Oakbnx>k. IL) in conjunction with a Sony high-resolution color monitor. Two epifluorescence filter

combinations were used: (a) 485 DF22 excitation/530 DF30 barrier/505 DRLPdichroic for observing the reduced form of the dye and (b) 633 DF10

excitation/675 DF30 barrier/650 DRLP dichroic for observing the oxidizedform of the dye. A Nikon 60X CFN Plan Apochromat oil immersion objective(numeric aperture 1.4) was used for fluorescence imaging. Subconfluent cellswere observed in custom made chambers which consisted of a Falcon 35-mmcell culture dish in which a 19-mni circle was drilled out and replaced with a25-mm square, no. I Corning cover glass. Cells were incubated for 30 min with

2 ml of HBSS containing dye (1 to 100 PM). washed free of excess dye. placedin 2 ml of growth media, and observed, or dye was added directly to cells andthe accumulation was examined in real time.

Eppendorf Centrifuge Tube

RESULTS

Physical and Photochemical Data. The molecular structure of thedyes used in this study along with pertinent physical and photochemical data is given in Table 1. The novel chromophores are planar,possess a delocalized positive charge, absorb red light efficiently, andare very lipophilic when compared to other common photosensitizers(8). For the present study we ethylated the 5-amino group of the

benzophenoxazine analogues because structure/function relationshipstudies had shown, unexpectedly, that this modification generatescompounds with greatly increased aqueous solubility even thoughthere is a concomitant increase in lipophilicity. Apparently, the increased steric hindrance resulting from the alkylation leads to a decreased tendency to form insoluble aggregates in aqueous solutions.As expected, a comparison of the absorption maxima shows that abathochromic shift results following the replacement of the oxygenatom in EtNBA with an atom of either sulfur or selenium (16). In lieuof directly measuring the triplet quantum yields of the new chromophores we have used their quantum efficiency for generating singlet oxygen as a qualitative measure of their potential for increasedphotoactivity. The quantum yields of 'O2 formation were determined

using the 1,3-dihenylisobenzofuran bleaching method, which is de

tailed elsewhere (6). Recently, Becker el al. (30) have shown a correlation between triplet quantum yield and singlet oxygen yield for aseries of closely related benzophenoxazinium compounds. The "heavyatom effect" substantially increased the singlet oxygen yields of

EtNBS and EtNBSe to 0.025 and 0.65, respectively, when comparedthe low yield of 0.005 of EtNBA.

Uptake and Retention of New Chromophores. Quantitation ofcellular dye accumulation revealed that the three analogues exhibitedsimilar uptake patterns (Fig. 2) with a rapid initial intracellular uptakefollowed by a more gradual increase in the ¡ntracellularconcentration.The plateau represents depletion of dye from the extracellular mediumrather than saturation of the cells sequestering capacity. This wasverified by the serial addition of dye to cells after the initial plateau inuptake was reached. This resulted in additional dye uptake and a new,higher plateau with each new addition (data not shown). At the end ofa 30-niin incubation period with 2 ml of 3 UMdye, the concentrations

of EtNBA, EtNBS, and EtNBSe were 2.72 ±0.10, 3.11 ±0.15, and2.94 ±0.03 (SD) nmol/106 cells, respectively. There is apparently no

direct correlation between dye uptake and the partition coefficient ofthese analogues since the analogue with the lowest partition coefficient, EtNBSe, accumulated to the same extent as the others. Underthe same conditions, EMT-6 cells accumulated 6.05 ±0.91 nmol/106

cells of PII. It should be mentioned that Lin et al. (IO) found that therapid uptake of benzophenoxazines was not significantly inhibited bythe presence of 10% serum in the extracellular medium, unlike the

NH(C,HJ (CI-)

Fhypo%«^

. Oil Layer

Cells withDye-^,Photodiode

Detectorg.1. Schematic of cell holder for recording the absorption spectrum of dyes under

xic conditions.S^Table

1 Physical and photochemical properties of pertinenti/vc.vDye

X'O2"EtNBA

O 0.005EtNBS S 0.025EtNBSe Se 0.650Amnx

<nm)*f632

65265965.800

68.60081.900"

Absolute quantum yields of 'O? formation determined by the diphenylisoben?

bleach method (Ref. 3).'' Measured in methanol containing 1% 1.0 M HC1.' Partition coefficients between 2-octanol and phosphate-buffered saline at pHPc'495

580120ofuran

7.4.

2573

Research. on August 20, 2021. © 1993 American Association for Cancercancerres.aacrjournals.org Downloaded from

PHOTOTOXICITY OF BENZOPHENOXAZ1NE ANALOGUES

3.5

UT= 3

"o 2.5

M* 2o

e 1.5

0.5

EtNBA

EINES

- -»- EtNBSe

1 5 30Time (min)

45

Fig. 2. Uptake curves of benzophenoxazine derivatives by EMT-6 ceiis. Cells (3 XICCY25cm2) were allowed (o attach overrate. Dyes (2 ml M 3 UM in phenol red-free HBSS)

were added to the cells, and cellular dye concentrations were determined at different timeintervals after incubation at 37°C. Dye in cells was extracted with acidified chloroform-

:methanol and quantitated spectrophotometrically. Points, means of 4 determinations:bars. SDs when they exceed symbol size.

case with PII. Likewise, we have found this to be true for these newderivatives (data not shown).

To determine whether the accumulation of the chromophores ¡sasaturable process (Fig. 3), dye uptake was studied by varying theextracellular dye concentration from 0.50 to 100 U.M.Only the 30-min

interval for incubation was examined. The results indicate that thecellular uptakes of EtNBA, EtNBS, and EtNBSe were a linear function of the dye concentration and that the cells possess an efficientmechanism for sequestering these chromophores.

In order to determine the rate of dye uptake and whether the dyeswere bound to the plasma membrane or distributed intracellularly, weexamined the cells with both fluorescence and light microscopy.While the fluorescence from EtNBA and EtNBS was sufficient to beobserved visually by microscopy, the weak fluorescence from EtNBSerequired the enhancement provided by a silicon intensified targetvideo camera for observation. The intracellular distribution of thebenzophenoxazine derivatives was dependent on the extracellular dyeconcentration. Staining the cells with 0.01 to 0.10 UMdye for 30 minresulted in a similar punctate fluorescence pattern in the perinuclearregion with each dye (Fig. 4/4). The pattern resembles the one displayed by other members of this class of compounds which wereidentified as lysosomes (11). At concentrations greater than 0.10 UM,the fluorescence usually became too intense to resolve into distinctstructures (Fig. 4ß);the dye could be seen in most cytomembranesand organdíes (mitochondria) but was not observed in the nucleus.Bright-field microscopy of cells which were exposed to concentra

tions of dye equal to or greater than 0. l UMrevealed nonfluorescent,blue particles located mainly in the perinuclear region. When a benzophenoxazine is added to the extracellular medium while simultaneously monitoring the red fluorescence in real time, one observes thenearly instantaneous intracellular localization of dye. This rapid trans-

location of colorant appears to be characteristic of benzophenoxazines(8, 10). Contrariwise, when PII is added to the cells one initiallyobserves an intense staining of the plasma membrane followed by aslower migration to the perinuclear region with a concomitant increasein cytoplasmic fluorescence. When using a filter system with allowsthe emission of all wavelengths greater than 520 nm, a yellow punctate fluorescence pattern is seen in the perinuclear region (Fig. 4C).The temporal dependent changes in the cellular distribution of porphyrin fluorescence have been reported (31-33) and have been as

cribed to differences in both the mechanism of internalization as well

as fluorescence quantum yields of monomeric and aggregated fractions. Histochemical and biochemical studies are currently under wayin order to determine the specific subcellular localizations as well asthe mechanism of uptake of these novel benzophenoxazines. It is ofinterest to note that thermally denatured cells (60°Cfor 20 min) take

up approximately 70% of the EtNBA levels found in live cells andlocalize it diffusely in the perinuclear region.

The dye efflux data (Fig. 5) reveal that with the EMT-6 cell line

approximately 40% of EtNBS and EtNBSe were retained after 6 h indye-free media, whereas only 20% of the EtNBA was retained during

the same time period. Dye recovered from the cells during the courseof the experiment was structurally unaltered when assayed by thin-

layer chromatography (silica gel:methylene chloride/methanol, 90:10)and UV/VIS spectroscopy, indicating that it was stable in the cellularmilieu. The efflux data for PII revealed a 55% loss of dye during thefirst hour of incubation in a dye-free medium, going from 6.05 ±0.91to 2.75 ±0.52 nmol/106 cells with minimal losses occurring during

the subsequent 5 h of incubation and washings. Similar results havebeen previously noted for a variety of cell lines and were attributed tothe fact that following a short incubation period a large portion of theaccumulated porphyrins were loosely bound to the plasma membrane(31,32,34-36).

Photocytotoxicity Studies. The photocytotoxic behavior of thenew derivatives, as well as of PII, was examined over the concentration range 0.01 to 1.0 UMand is displayed in Fig. 6A. At a light doseof 3.3 J/cm2 an EDM)for cell killing occurred with sensitizer concen

trations of approximately 10.0, 0.10, and 0.01 UMfor PII, EtNBS, andEtNBSe, respectively. At the same light dose, 0.05 UMPII showed nophototoxicity, while EtNBSe killed 90% of the cells at this concentration. There is a correlation between the ability of the benzophenoxazine dyes to generate the cytotoxin 'O2 and their phototoxicity:

EtNBSe (0.65) > EtNBS (0.025) > EtNBA (0.005). However, theratios of ED50s do not correspond to the ratios of 'O2 yields. Fig. 6B

presents the survival kinetics of dye-treated cells as a function ofexposure to red light (0.66 to 3.3 J/cm2). At an extracellular dye

concentration of 0.5 UM,the light dose required to kill approximately50% of the cells was 2.0 and 0.5 J/cm2 for the S and Se analogues,

respectively. The EDM>for both EtNBA and PII could not be determined under these conditions because of their negligible phototoxiceffects. EMT-6 cells treated with 1.0 ml of HBSS containing 1.0 UM

dye for 30 min in the dark exhibited the following order of viability:PII (100%); EtNBA (86%); EtNBSe (78%); and EtNBS (61%). Thedark toxicity was reduced dramatically with 0.5 UMdye concentrations

= 100-o

M» 1 0-oEe

1 -

0.1

1 0Dye In Medium (uM)

100

Fig. 3. Effect of dye concentration on the uptake of benzophenoxazines by EMT-6

cells. Experiments were performed as described in Fig. 2 except with varying dye concentrations (0.5 to IOO pM) in the medium and at only the 30-min incubation interval.

Points, means of 4 determinations; SDs were less than 10%.

2574

Research. on August 20, 2021. © 1993 American Association for Cancercancerres.aacrjournals.org Downloaded from

PHOTOTOXICITY OF BENZOPHENOXAZINE ANALOGUES

Fig. 4. Fluorescence micrographs of EtNBA. Nile red, Meldola's blue, and PII in EMT-6 cells. Photographs were (aken wilh a Nikon N6006 camera and Kodak Ekiachrome 400

slide film. In u. after a 30-min incubation with 0.10 UMEiNBA. the dye appears in a punctate fluorescence pattern in the perinuclear region which is characteristic of lysosomes. Inb, with 1.0 UMEtNBA. the fluorescence appeared Ux>intense to resolve into distinct cellular structures; however, the dye did not appear to stain the nucleus. In (. wilh 1.0 pvi of PIIand a filter which allows the observation of all emissions past 520 nm. the porphyrin is concentrated in the plasma membrane and cytoplasm (red), w'hile a yellou punctate fluorescente

pattern (unknown) appears in (he perinuclear region. In d. with 10.0 UMEtNBA and the reducianl sodium dithionite. the fluorescence from Ine reduced form of the dye appears in punctatevacuoles throughout the cell as well as in cytomembranes. In p (10.0 UMNile red) a neutral, lipid-spccific benzophenoxazine is shown for comparison with J. In/ 10.0 UMMeldola's

blue, a benzophenoxazine which can exist only in a charged form, also appears to concentrate in punctate vacuoles and cytomembranes but not the nucleus

(percentage viability): PII (100%); EtNBA (100%); EtNBSe (94%);and EtNBS (90%). Red light had no measurable cytotoxic effect onuntreated cells.

Cellular and Solution Reduction of Dye. A series of reflectancespectra obtained over a 30-min period from freshly isolated, stained

cells under anoxic conditions verified that the intracellular bleachingof the dyes does occur. Fig. 7 is representative of the result obtainedfor all three dyes. Although quantitative differences in the rates ofreduction between the various chromophores are difficult to obtain bythis procedure, they appear to be linear for all three dyes during thefirst 15 min of observation. The reversible redox behavior of the dyeswas demonstrated in anaerobic, decolorized cells by the introductionof air and observing their rapid recolorization. The enzymatic natureof the reduction was made apparent by cooling stained cells to 0°C,

thereby lowering enzyme reaction rate constants and thus shifting the

redox equilibrium toward the oxidized, blue form of the dye (Fig. 8).On warming the cells to 37°Cthe equilibrium is shifted in favor of thecolorless reduced form, while heating the cells further to 60°Cdena

tures the enzymes, whereupon the cells remain in the blue oxidizedform under all circumstances.

We utilized a cell-free system in order to further investigate the

propensity of the benzophenoxazines to undergo enzymatic reductionin the dark as well as photoreduction with red light. NADH waschosen as a reductant for the system not only because of its importance as a cofactor for many cellular reactions but also as a result ofthe extensive literature pertaining to its role in intracellular dye reductions (17-22). In order to increase the rate of reduction in the

presence of NADH, the widely distributed hydride transfer enzymediaphorase (17, 37-39) was also included in the system. During a5-min observation period in the dark and in the absence of oxygen.

2575

Research. on August 20, 2021. © 1993 American Association for Cancercancerres.aacrjournals.org Downloaded from

PHOTOTOXiriTY OF BENZOPHBNOXAZINB ANALOGUES

- 3.5

Fig. 5. Efflux pallems of ben/ophenoxazine derivatives from dye-loaded EMT-6 cells.Cells (3 x IO*/25 cm2) were incubaied wiih 2 ml of HBSS containing 3 MMdye for 30 minat 37°Cto allow uptake of dye. The cells were washed free of dye and overlayed with

RPMI 1640. Cellular dye concentrations were determined at various time intervals as forFig. 2. Points, means of 4 determinations; bars. SDs when they exceed symbol size.

rapidly. The addition of dithionite to unstained cells caused nonoticeable increase in this fluorescence pattern.

DISCUSSION

Our ability to easily manipulate the photophysical properties of thebenzophenoxazines through structural modifications has led to increased triplet state quantum yields in both EtNBS and EtNBSe whencompared to EtNBA and has greatly enhanced their photochemother-apcutic efficacy in EMT-6 cells. This was especially true for EtNBSe,which has a 'O2 yield of 0.65. In fact, under comparable incubation

conditions, EtNBSe is 3 orders of magnitude more phototoxic thanPII, requiring only 0.01 UMdye to kill 50% of the cells. Lin et al. (IO),using the MGH-U1 human bladder carcinoma cell line, have alsoreported a 1000-fold greater potency of a benzophenoxazine analogue

compared to HPD. The large difference in photosensitizing abilitybetween these two classes of dyes is most likely a result of a combination of factors. First, the molar extinction coefficient of PII inmethanol at 625 nm is only 2,300 litcrs/mol/cm compared to approxi-

NADH alone reduced 4% and 10% of EtNBS and EtNBSe, respectively, with no noticeable reduction of EtNBA. Fig. 9A indicates thatwhen diaphorase was combined with NADH under these same conditions both EtNBS and EtNBSe were reduced by approximately 50%in the first 5 min (í,/2),while the extrapolated tV2 for EtNBA was 12min. Diaphorase alone caused no bleaching of the dyes. Aeration ofthe bleached samples resulted in substantial reoxidation of the dyes tothe colored form (~70%), thus demonstrating the reversibility of the

reaction in the presence of oxygen. Prevention of total recovery isprobably due to both precipitation of the highly insoluble reducedform as well as to some adsorption of the oxidized form to the glasscuvet.

Irradiation of the chromophores with red light (590-700 nm, 0.048W/cm2) in the presence of only NADH (Fig. 9B) and in the absence

of oxygen caused a dramatic photoreduction of EtNBSe and EtNBSleading to /l/2s of 30 s and 2 min, respectively. EtNBA appeared stableduring the time period studied. As above, aeration of the photoreduceddyes leads to substantial reoxidation to the colored form. In the timeperiod studied, there was no appreciable reduction of dye when theirradiation was carried out in the presence of oxygen. This indicatesthat under realistic (therapeutic) conditions the dyes will be efficacious.

Video-enhanced fluorescence microscopy of dye-treated cells using

the filter system specific for observing the reduced form of the dyes(green emission) revealed their localization in cytomembrane structures as well as in punctate vacuoles throughout the cytoplasmicregion (Fig. 4D). The degree of fluorescence varied from experimentto experiment and is most likely a result of cell density and themetabolic state of the cells at the time of observation. Under the sameconditions Nile red (Fig. IO), a lipid-specific stain (40, 41), gave a

very similar fluorescence pattern (Fig. 4£).It is noteworthy that Gutz(42) reports that the staining of plant cells with either reduced Nileblue or Nile red always caused fluorescence of the same lipid-rich

structures. Verification that the fluorescence is a consequence of thereduced form of the dye was made by adding an exogenous reductant,sodium dithionite, to the cells and observing the concomitant increasein green fluorescence with the simultaneous disappearance of the redfluorescence from the oxidized form of the dye (not shown). One alsoobserves the same fluorescence patterns if the dyes are reduced withdithionite prior to addition to the cells. The autofluorescence observedwith this filter system was minimal and had a tendency to bleach

2576

J>x

3CO

•¿�o

1 0-

1-

0.1

f X

0.01Dye

0.1Concentration (uM)

100-

_>>

(O

o

0.1

mI-.—

-*—¿�M-

--i--,..p"

p*..IEtNBA*»JEtNBS

r».EtNBSe*»,ìb

0.5 1.5 2.5 3.5

Dose ( Jle m \Fig. 6. The pholocyloloxicily of PII and benzophenoxazine derivatives in EMT-6 cells.

In a. cells (0.5-2.0 x I0"/0.5 ml) were seeded in a 24-well. flat-bottomed Linbro plates.

Forty-eight h later, subconfluent monolayers were incubaied for 30 min in the dark withI ml HBSS containing I.O. 0.5. 0.1. 0.05. or 0.01 UM sensitizer at 37°C. Following

incubation, all cultures within each group were washed twice with HBSS and subjected toirradiation (IO min: 3.3 J/cm2; 590-700 nm). In h. the same procedure was followed as

with a, except that the cells were treated with 0.5 UMdye and irradiated with 0.66, 1.32.1.98. 2.64. or 3.3 ¡/cm2.After treatment all cultures were placed back into growth mediaat 37°Cuntil assayed for cell viability 48 h following irradiation via ['HJthymidine

incorporation. Ptiints. means of 6 determinations; bars. SDs when they exceed symbolsize.

Research. on August 20, 2021. © 1993 American Association for Cancercancerres.aacrjournals.org Downloaded from

I'HOTOTOXiriTY Oh HI N/OIMII \()\ \/IM XNMcXilï.S

1.0 _

0.8 -

CDttO(/>00

0.2 -

0.0

400 500 600WAVELENGTH

700 800

Fig. 7. The reduction of ElNBA by F.MT-6 cells. Cells (3 X l()fi/25 cm2) were allowed to altach overrule. ElNBA (2 ml at 3 UMin phenol red-free HBSS) was added to the cellsand incubated at 37°Cfor 30 min in the dark, washed with HBSS. trypsini/ed. and placed in clear 1.5-ml microcentrifuge tubes. The cells were pellett/ed by centrifugaron ( 10.000

x £)for 2 min. the HBSS and trypsin were removed by aspiration, and the cells were layered with 0.5 ml of silicone oil. Recordings of the absorption spectrum were made every 5min using reflective absorption spectromelry.

mately 70,000 tor the ben/.ophenoxazines. More importantly, integration of their corresponding absorption curves shows that the ben-

/ophenoxazines absorb 30 times more red light than PII. Second, the30-min incubation time used in these studies restricts a large portionof the porphyrin to the plasma membrane (34, 43^15), where photo-

dynamic damage may be less severe or repair mechanisms moreefficient. During this same time period, fluorescence microscopy indicated that the benzophenoxazines were rapidly accumulated intointracellular organdíes where photodamage may be much more lethal.While we did not measure the 'O2 yield of PII. reported values vary

between 0.13 and 0.85. depending on the method and media used forthe measurement (46. 47). This suggests that 'O: yield is not a major

contributing factor for the large differences observed in photokillingbetween the two classes of dyes.

Although we are currently examining the mechanisms involvedwith uptake/retention and intracellular localization of the dyes, thereare some preliminary considerations which are worthy of discussion at

this time. Our observation of the near-instantaneous entry of the

chromogens into the cells coupled with those by Lin et al. (10) andBastos el al. (48) indicating that the uptake of benzophenoxazinesproceeds at low temperatures seems to preclude endocytic or pinocyticmechanisms. Significantly, as Fig. 8 illustrates, these chromophoresreadily undergo a protonation/deprotonation reaction that results in aneutral imino compound. Many studies have shown that compoundswhich exhibit this behavior often have increased rates of entry into thecell (49-51). Since the pKa values of the three new dyes are approx

imately 11.0 in water, one can expect low extracellular concentrationsof the neutral/deprotonated species at physiological pH. Therefore, theneutral form would have to be highly membrane permeable in order toaccount for the rapid ¡ntracellularaccumulation. The ease of penetration of the plasma membrane by a neutral benzophenoxazine wasdemonstrated with Nile red (Fig. 4£),a dye that is structurally verysimilar to the neutral imino compounds (Fig. 10). Thus, the first stepin cellular sequestration may consist of an energy-independent, simple

Fig. 8. Redox and protonalion/deprotonationequilibrium associated with bcn/.ophenoxa/ine analogues (X = O, S, Se).

(C2H5)2N

Lightor

Enzyme

(C2H5)2N

NeutralColorless

2577

MNH(C2H5)

Cationic(630-660nm)

Oxygen

NH(C2H5) (C2H5)2N N(C2H5)

Neutral(520-530nm)

Research. on August 20, 2021. © 1993 American Association for Cancercancerres.aacrjournals.org Downloaded from

l'HOÕOTOXU'lTY B[.N/()I'H1.N()\A/[M \V\UXUTS

diffusion of the neutral form of the dye into and across the lipidplasma membrane with subsequent equilibration, with the protonatedform occurring in the intracellular milieu.

The possibility that the cationic form may simply diffuse across themembrane in response to either the negative plasma membrane potential or an electrostatic attraction to cellular polyanions must also beconsidered. The argument that charged molecules are inhibited fromcrossing membranes most often applies to compounds possessing alocalized point charge. It is questionable whether this associationapplies to lipophilic compounds possessing a delocalized charge suchas the benzophenoxazines. Benz (52). using a series of tetraphenyl-

borate analogues which exist exclusively in a cationic form, hasshown that the ability of lipophilic ions to cross membranes is directlyrelated to the area over which the charge is distributed. The greater thedelocalization, the easier it is to translocate the membrane. This behavior is observed with many lipophilic ions (Refs. 52 and 53 andreferences therein) including Meldola's blue (Fig. IO), a positively

charged benzophenoxazine, which is unable to form a neutral speciesand yet (we found) still permeates the plasma membrane of EMT-6

cells at a rate comparable to that of EtNBA (see Fig. 4F). It appears.

100

(C2H5)2N

75-

>>O

3 50-1

KO

* 25-I

-*— EtNBA+Dph+NADH-• —¿�EtNBS+Dph+NADH-•—¿�EtNBSe+Dph+NADH

-A- - EtNBA-fNADH

-a - EtNBS+NADH

- •¿�- EtNBSe+NADH

2 3Time (min)

75-•

>>•o

S50-•oXO*

25-•C=

—¿�»-=----*^-:ï-- +~ f= r= t —¿�S—¿�SO^â„¢¿�~' ^ ^ _^__^X

K —¿�EtNBA+Dark4 'k ---»- EtNBS+Dark

X —¿�•¿�—¿�EtNBSe+Dark«

\ A—¿� EtNBA+Light

*. °v —¿�-B —¿�EtNBS+Light

% X - - -0 •¿�EtNBSe+Light1

' 1 ' 1 ' 1

1234Time

(min)-

—¿�ib5

Fig. 9. Enzymatic and photoreduction of benzophenoxazines in a cell-free system. Ina, to a cuvet equipped with a septum cap was added 2.0 ml of dye ( IO UMin HBSS) and0.2 ml of NADH solution ( I mg/ml HBSS). The solution was deaerated by purging withargon gas in the dark for 15 min. After the injection of 0.2 ml of argon-purged diaphorase(Dph) solution ( 1 mg/ml in HBSS) the disappearance of dye was monitored speclropho-tometrically at room temperature at 1-min intervals. The independent effect of NADH

with dye was also measured under the same conditions. In b, to a cuvet equipped with aseptum cap was added 2.0 ml of dye ( 10 UMin HBSS) und 0.3 ml of NADH solution(1 mg/12 ml HBSS). The solution was purged with argon gas in the dark for 15 min.Absorption readings at the dye's maximum were taken every minute for 5 min in the dark,

and then the solution was exposed to red light (590-700 nm, 0.048 W/cm2), and absorp

tion readings were taken every 30 s {5 s for EtNBSe) for an additional 5 min.

CI

Meldola's BlueFig. 10. Structures of Nile red and Meldola's blue.

therefore, that either of the two possible forms of the dye can readilycross into cells, thus accounting for the fact that various benzophenoxazines having diverse pKa's all seem to accumulate intracellularly

at about the same rate (10).Several accepted mechanisms for rationalizing the intracellular lo

calization patterns of the benzophenoxazines are also based on the twophysicochemical characteristics which appear to facilitate their rapidtraversal of the plasma membrane, i.e., their lipophilic nature andability to exist either as a charged or neutral molecule. The partitioning into cellular membranes and the extended retention times observedfor these chromophores are most likely the result of hydrophobicinteractions of the dyes with the lipophilic membrane constituents(lipoproteins, phospholipids, etc.) in addition to electrostatic interactions between the positively charged dyes and the negatively chargedcomponents of the membranes. While we have yet to identify thefluorescent punctate structures in the perinuclear region as lysosomes,Lin et al. (11), using a variety of similar benzophenoxazines, havedone so by histochemical methods. The accumulation of weakly basicdrugs into acidic compartments, such as lysosomes, is generally explained by an "ion trapping" mechanism (51, 54). When compounds

that can undergo protonation/deprotonation conversions encounter amembrane-delimited acidic body, the neutral/deprotonated form dif

fuses across the membrane into the acidic interior where it becomesprotonated and presumably cannot readily diffuse out. Finally, themitochondrial localization of other benzophenoxazine derivatives hasbeen observed previously (II) and identified cytochemically by Bastos and Marques (48) with a double staining technique for succinicdehydrogenase activity. While it is often proposed that the negativeelectrical potential across mitochondrial membranes is primarily responsible for the accumulation of positively charged dyes (55, 56), wehave shown that this is not the case with benzophenoxazines (11) byexamining dye uptake in the presence of agents known to reduce themitochondrial membrane potential. None of the treatments causedsubstantial reduction in dye uptake. On the other hand the staining ofmitochondria by lipophilic benzophenoxazine dyes may simply be dueto a combination of electrostatic and hydrophobic interactions with thenegatively charged (cardiolipin) mitochondrial membranes (57). Theimportance of hydrophobic interactions between dyes and mitochondria has been demonstrated with a variety of cationic chromophoreswhich initially do not stain mitochondria but upon minor structuralchanges to increase their lipophilicity become preferentially bound(50, 58, 59). It should be mentioned that the observed reduction of thedyes necessitates that they come in contact with reductases/

2578

Research. on August 20, 2021. © 1993 American Association for Cancercancerres.aacrjournals.org Downloaded from

PHOTOTOXICITY OK BliNZOPHF.NOXA/JNK ANALOGUES

dehydrogenases and their coenzymes. NAD(P)H. The cellular locations most likely to contain these types of enzymes/coenzymes are themitochondria, the cytosol. and the endoplasmic reticulum. implyingthat either the dyes are localized to some extent at these sites or anequilibrium exists between the dye in these organdíes and otherintracellular reservoirs.

A significant aspect of this study is the finding that this class ofchromophores is capable of undergoing an intracellular. reversibleenzymatic reduction. While previous studies have indicated that bacteria, fungi, and plant cells are capable of reducing benzophenox-

azines (20, 42, 60), to the best of our knowledge this is the firstmeasurement of its occurrence in a mammalian cell line. The importance of this observation cannot be overemphasized since the ratio ofoxidized to reduced dye will contribute profoundly to dye retention,localization, and photodynamic efficacy. Not only does the enzymaticreduction of the benzophenoxazines give a neutral species (Fig. 8), italso results in a colorless compound which no longer absorbs light inthe "therapeutic window" and therefore is not phototoxic. Without

knowing the degree to which this reduction occurs in monolayeredcells cultured under normal oxygen tension, one cannot quantitativelyassess correlations between dye-mediated cell killing and 'O2 yields.

This intracellular reduction may explain past results (6, 10) whereother benzophenoxazines having high 'O2 yields had unexpected low

photokilling. The ease of reduction of a related class of dyes known asphenothiazines (méthylèneblue, thionin, toluidine blue. Azure A. etc.)is well known (17. 19, 61-63) but does not seem to be generally

appreciated in the area of PDT even though it may explain the generally poor phototoxic results observed with these dyes both in vitroand in vivo. Our studies have shown that even in the presence ofoxygen the enzymatic reduction of these phenothiazines to their colorless leuko form is substantial (data not shown). However, it isevident from the enhanced cytotoxic behavior of EtNBS and EtNBSethat under normal in vitro conditions a substantial portion of thesedyes exists in the oxidized form. This is easily verified by the blueappearance of the monolayered cells. It remains to be seen whether thein vitro results can be translated to solid tumors. Whether the hypoxicareas of the tumor will contain enough oxygen to provide a therapeuticlevel of oxidized dye is a major concern.

The potential for redox cycling during irradiation (i.e., photoreduc-

tion; see Fig. 9B) may influence the phototoxic efficacy of these dyesin the following two distinct and opposite ways. Photoreduction decreases the concentration of oxidized dye available for the generationof the cytotoxin 'O2 (type II mechanism), therefore decreasing PDT

effects. However, photoreduction may result in the direct abstractionof electrons from important biomolecules with the formation of freeradicals and oxidants (type I mechanism) therefore increasing PDTeffects. Oxidants derived via type I mechanisms have been detectedduring the photoreduction of phenothiazines and acridine dyes in thepresence of NADH ( 19, 64. 65). Obviously, the net effect of photoreduction on PDT efficacy will be determined by a complex interactionof these opposing activities. The reversible photoreduction observedin the presence of oxygen with the benzophenoxazine class of dyesshould not be confused with the irreversible "photobleaching" of

porphyrin type photosensitizers that is observed during PDT (66, 67).It is believed that these photosensitizers. unlike the benzophenoxi-

zines, are destroyed during PDT by an irreversible photooxidativeprocess.

Studies are currently under way to delineate more precisely howintracellular reduction and protonation/deprotonation affect dye localization and cytotoxicity in vitm and in vivo. Preliminary data in ourlaboratory indicate that a rapid and extensive cellular uptake of thedyes occurs in vivo, thus differentiating these dyes from PII. whichlocalizes primarily to the tumor stroma. It will be of interest to see if

the enhanced cell killing observed with EtNBS and EtNBSe whencompared to PII will translate to the animal model system.

ACKNOWLEDGMENTS

The authors wish to thank Dr. Alan R. Oseroff for his valuable suggestionsand criticisms of the manuscript; Monica Gome/. Tracy MacEachern, andDonald Rogers for technical assistance: Jay Scarpelli and MaryAnn Nilsson fortheir photographic assistance; and Nik Kollius for use of the reflective absorption spectrometer.

REFERENCES1. Dougherty. T. J. Photosensitizers: therapy and detection of malignant tumors. Photo-

chem. Photobiol.. 45: 879-889. 1987.2. Henderson. B. W.. and Dougherty. T. J. How does photodynamic therapy work?

Photochcm. Photobiol.. SS: 145-157. 1992.3. Lin. C-W. Pholodynamic therapy of malignant tumors: recent developments. Cancer

Cells (Cold Spring Harbor). J: 437-144. 1992.4. Rosenthal. I Phthalocyanines as photodynamic sensitizers. Photochem. Photohiol..

53: 859-870. 1991.

5. Corner, C. J. Preclinical examination of first and second generation photosensitizersused in photodynamic therapy. Photochem. Photohiol.. 54: 1093-1107. 1991.

6. Cincona. L., Foley. J. W.. and Cincona. A. H. Novel red absorbing bcnzo[¿/|phenox-.1/1nMI:n and benzol«Iphenothiazinium pholosensitizers: in vitro evaluation. Pholo-ehem. Photobiol.. 46: 751-758. 1987.

7. Foley. J. W., Cincona. L., and Cincona. A. H. Structures and properties of novelbenzo|«lphenothiazinium phoiochcmotherapcutic agents. SPIE Proc.. 847: 90-95.

1987.8. Cincona. L.. Foley. J. W.. and Cincona. A. H. Novel phenothiazinium photosensitizers

for photodynamic therapy. SPIE Prix:.. 997: 145-153, 1988.9. Cincona. A. H.. Cincona, L., and Foley. J. W. Novel benzophenothiazinium photo

sensitizers: preliminary m vin. results. SPIE Proc.. 1203: 202-210. 1990.10. Lin. C-W.. Shulok. J. R.. Wong. Y-K.. Schanbacher. C. F.. Cincona. L., and Foley. J.

W. Photosensitization, uptake, and retention of phenoxazine Nile blue derivatives inhuman bladder carcinoma cells. Cancer Res.. 51: 1109-1116. 1991.

11. Lin. C-W.. Shulok. J. R.. Kirley, S. D.. Cincona. L., and Foley. J. W. Lysosomal

localization and mechanism of uptake of Nile blue pholosensitizers in lumor cells.Cancer Res., 51: 2710-2719. 1991.

12. Lewis, M. R., Sloviter. N. A., and Goland. P. P. In viro staining and retardation ofgrowth of sarcomata in mice. Anal. Ree.. 95.- 89-96. 1946.

13. Lewis. M. R.. Goland. P. P.. and Sloviter. N. A. The action of oxazine dyes on tumorsin mice. Cancer Res., 9: 736-740, 1949.

14. Drexhage. K. H. Structure and properties of laser dyes. In: F. P. Schafer (ed.). Topicsin Applied Physics. Vol. I. pp. 144-193. New York: Springer. 1977.

15. Spikes, J. D., and Straight, R. Sensitized photochemical prwesses in biologicalsystems. Annu. Rev. Phys. Chem., 18: 409-436, 1967.

16. Deny. M. R., and Merkel. P. B. Chalcogenapyrylium dyes as potential photochemo-therapeutic agents. Solution studies of heavy atom effects on triplet yields, quantumefficiencies of singlet oxygen generation, rates of reaction with singlet oxygen, andemission quantum yields. J. Am. Chem. Soc., 112: 3845-3855, 1990.

17. Benhetto, H. P.. and Stirling. J. L. Reduction of redox mediators by NADH andelectron transduction in bioelectrochemical systems. Chem. Ind. (Lond.). 20: 695-697, 1985.

18. Zbaida, S., and Levine. W. G. Role of electronic factors in binding and reduction ofazo dyes by hepatic microsomes. J. Pharmacol. Exp. Ther., 260: 554-561. 1992.

19. Julliard. M.. and Le Petit. J. Regeneration of NAD* and NADP4 cofaclors by

photosensitized electron transfer. Photochem. Photobiol.. 36: 283-290.20. Gots, J. S.. Jordan. V. E.. and Brodie, A. F. Studies on the action of nitrofurans on

bacterial enzyme systems. III. furacin interference with dye reductions by Esi'herii-huicoli. Arch. Biochem. Biophys.. 36: 285-298, 1952.

21. Muller. P. K.. Krohn. K.. and Muhlradt. P. F. Effects of pyocyanine. a phenazine dyefrom Pseudónimas aeruginosa. on oxidative burst and bacterial killing in humanneutrophils. Infect. Immun.. 57: 2591-2596, 1989.

22. Hunter. J. C.. Woods. M. W.. and Burk. D. Inhibition of anaerobic glyeolysis of ascitescancer cells by acridine dyes and light. J. Nati. Cancer Ins!.. .<4: 535-541. 1965.

23. Groves, J. T., Haywood. B. J.. and Knol. J. A. Synthesis of selenotoluidine blue. J.Med. Chem.. 17: 902-904. 1974.

24. Crossley, M. L., Dreisbach, P. F.. Hofmann. C. M., and Parker, R. P. Chemothera-peutic dyes. I. 5-aralkylamino-9-alkylaminobenzo|u|phenoxazines. J. Am. Chem.Soc., 74: 573-577, 1952.

25. Clapp, R. C.. English, J. P.. Catherine. E. F. Forsythe. J., Grolz, R. E., and Shepherd.R. G. Chemotherapeutic dyes. V. benzo|«]phenothiazines and benzo[u|phenazines. J.Am. Chem. Soc., 74: 1994-1996. 1952.

26. Rockwell. S. C.. Kallman. R. F.. and Fajordo, L. F. Characteristics of a seriallytransplanted mouse mammary tumor and ¡istissue culture adapted derivative. J. Nati.Cancer Inst., 49: 735-749. 1972.

27. Rockwell. S. C.. and Kallman. R. F. Cellular radiosensitivity and tumor radiationresponse in the EMT6 tumor cell system. Radial. Res., 53: 281-294. 1973.

28. Frcshney, R. I. Cylotoxicily and viability assays. In: D. Rickwood and B. D. Harnesteds.). Animal Cell Culture, pp. 183-216. Oxford. England: IRL Press. 1986.

29. Richter. A. M.. Waierfield. E.. Jain. A. K.. Slernberg. E. D.. Dolphin. D., and Levy.J. G. In vitro evaluation of phototoxic properties of four structurally relaled ben-

2579

Research. on August 20, 2021. © 1993 American Association for Cancercancerres.aacrjournals.org Downloaded from

PHOTOTOXICITY OF BENZOPHENOXAZINE ANALOGUES

zoporphyrin derivative*. Photochem. Pholobiol.. 52: 495-500, 1990.

30. Becker. R. S.. Chaknivorti, S.. and Das. S. The photosensitizeâ„¢ benzophenoxazineand thia/ines: comprehensive investigation of photophysical and photochemical properties. Pholochem. Pholobiol.. 51: 533-538, 1990.

31. Bottiroli. G.. Croce. A. C.. Ramponi. R., and Vaghi. P. Distribution of disulfonatedaluminium phthalocyanine and photofrin II in living cells: a comparative fluorometricstudy. Photochem. Photobiol., 55: 575-585. 1992.

32. Xiao, Y. H. and Jacques. S. L. Effectiveness of photosensitive dye during uptake andredistribution. SPIE Proc.. 1645: 205-215. 1992.

33. Smith, G. J. The effects of aggregation on the fluorescence and the triplet state yieldof hematoporphyrin. Photochem. Pholobiol.. 41: 123-126, 1985.

34. Kessel. D. Sites of photosensitization by derivatives of hematoporphyrin. Photochem.Photobiol.. 44: 489-494. 1986.

35. Boegheim. J. P. J.. Lagerberg. J. W. M.. Tijssen, K.. Dubbelman, T. M. A. R., andSteveninck, J. V. Preferential uptake of cytotoxic porphyrins from hematoporphyrinderivative in murine L929 fibroblasts and Chinese hamster ovary KI epithelial cells.Biochim. Biophys. Acta, IOI2: 237-242. 1989.

36. Kessel. D., and Chou. T-H. Tumor localizing components of the porphyrin preparationhematoporphyrin derivative. Cancer Res., 43: 1994-1999, 1983.

37. Savage. N. Preparation and properties of highly purified diaphorase. Biochem. J., 67:146-155, 1957.

38. Culling. C. F. A., Allison, R. T.. and Barr. W. T. In: Cellular Pathology Technique. Ed.4, pp. 326-327. London: Butterworths. 1985.

39. Siegel, D.. Beali. H.. Senekowitsch. C.. Kasai. M.. Arai. H., Gibson, N. W., and Ross,D. Bioreductive activation of mitomycin C by DT-diaphorase. Biochemistry, 31:7879-7885. 1992.

40. Greenspan. P.. and Fowler. S. D. Spectrofluoromelric studies of the lipid probe. Nilered. J. Lipid Res.. 26: 781-789. 1985.

41. Fowler. S. D.. and Greenspan. P. Application of Nile red, a fluorescent hydrophobicprobe, for the detection of neutral lipid deposits in tissue sections. J. Histochem.Cytochem.. 33: 833-836. 1985.

42. Gutz. H. Fluorescent staining with Nile blue of the granula of the hyphae of Mucorracemosus. Planta (Beri.). 46: 481-511. 1956.

43. Christensen. T.. Sandquest. T.. Feren, K.. Waksvik. H.. and Moan. J. Retention andphotodynamic effect of haematoporphyrin derivatives after prolonged cultivation inthe presence of porphyrin. Br. J. Cancer. 48: 35-Õ3. 1983.

44. Moan, J.. McGhie, J., and Jacobsen. P. B. Photodynamic effects on cells exposed tohematoporphyrin derivative and light. Photochem. Photobiol.. J7: 599-604. 1983.

45. Kessel. D. Effects of photoactivated porphyrins at the cell surface of leukemia LI2IO. cells. Biochemistry, 16: 3443-3449. 1977.

46. Gottfried. V.. Peled. D.. Winkelman. J. W.. and Kimel. S. Photosensitizers in organized media: singlet oxygen production and spectral properties. Photochem. Photobiol.,48: 157-163, 1988.

47. West, C. M. L.. and Moore. J. V. The photodynamic effects of photofrin II. hematoporphyrin derivative, hematoporphyrin. and tetrasodium-meso-tetra(4-sulfonatophe-

nyhporphine in vilnr. clonogenic cell survival and drug uptake studies. Photochem.Photobiol.. V9: 169-174. 1989.

48. Bastos. A. L.. and Marques. D. Nile blue inducible fluorescence of tumor cells. Z.Natursforsch. Teil B Anorg. Chem. Org. Chem., 27: 1395-1398. 1972.

49. Henderson. B. W.. Waldow. S. M.. Mang. T. S., Potter. W. R.. Malone. P. B., andDougherty. T. J. Tumor destruction and kinetics of tumor cell death in two experi

mental mouse tumors following photodynamic therapy. Cancer Res., 45: 572-576.

1985.50. Roding. J.. Naujok, A., and Zimmerman, H. W. Effects of ethidium bromide, tetram-

ethyl ethidium bromide and betaine B on the ultra structure of HeLa cell mitochondriain situ: a comparative binding study. Hislochemistry. #5: 215-222. 1986.

51. Goldstein. A.. Aronow. L.. and Kaiman. S. M. The absorption, distribution, andelimination of drugs. In: Principles of Drug Action: The Basis of Pharmacology, pp.129-225. New York: John Wiley and Sons, 1974.

52. Benz. R. Structural requirements for the rapid movement of charged panicles acrossmembranes: experiments with tetraphenylborate analogues. Biophys. J., 54: 25-33,

1988.53. Gros. P.. Talbot, F.. Tang-Wai, D., Bibi, E., and Kaback, H. R. Lipophilic cations: a

group model substrate for the multidrug resistance transporter. Biochemistry. 31:1992-1998, 1992.

54. deDuve, C.. de Barsy, T.. Poole, B.. Trouel, A., Tulkens. P., and VanHoff, F. Lysos-omotropic agents. Biochem. Pharmacol., 23: 2495-2531. 1974.

55. Johnson, L. V.. Walsh, M. L., and Chen. L. B. Localization of mitochondria in livingcells with rhodamine 123. Proc. Nati. Acad. Sci. USA, 77: 990-994, 1980.

56. Davis. S., Weiss. M. J., Wong. J. R., Lampidis, T. J., and Chen, L. B. Mitochondria!and plasma membrane potentials cause unusual accumulation and retention ofrhodamine 123 by human breast adenocarcinoma derived MCF-7 ceils. J. Biol.Chem.. 260: 13844-13850. 1985.

57. Huart. P.. Brasseur. R.. Gormaghtigh, E.. and Ruysschaert. J. M. Antimitotics inducecardiolipin cluster formation, possible role in mitochondrial enzyme inactivation.Biochim. Biophys. Acta. 799: 199-202. 1984.

58. Zimmermann. H. W. Physicochemical and cvtochemical investigations on the bindingof ethidium and acridine dyes to DNA and organdíes in living cells. Angew. Chem.Int. Ed. Engl.. 25: 115-130, 1986.

59. Brenner, S. Supravital staining of mitochondria with phenosafranin dyes. Biochim.Biophys. Acta. //: 480-486. 1953.

60. Mix. M. Mechanism of spherosome fluorescence with oxazines and other basic dyes.Protoplasma, 50: 434-470. 1959.

61. Buskirk. H. H.. Crim. J. A.. Van Giessen, G. J.. and Petering, H. G. Rapid in vivomethod for determining cytotoxicity of antitumor agents. J. Nati. Cancer Inst., 57:135-138. 1973.

62. Bellin. J. S., Mohos. S. C.. and Oster. G. Dye-sensitized photoinactivation of tumorcells in vitro. Cancer Res., 21: 1365-1371, 1961.

63. Barbosa, P., and Peters, T. M. The effects of vital dyes on living organisms withspecial reference to méthylèneblue and neutral red. Histochem. J., 3: 71-93, 1971.

64. Hunter, J. C., Bark, D.. and Woods. M. W. Influence of diphosphopyridine nucleotide(DPN) on phtodynamic effects of low concentrations of méthylèneblue in ascitestumor cells. J. Nail. Cancer Inst., 39: 587-593, 1967.

65. Martin. J. P.. and Naomi, L. Oxygen radicals mediate cell inactivation by acridinedyes, fluorescein and lucifer yellow CH. Photochem. Photobiol., 46: 45-53, 1987.

66. Mang. T. S.. Dougherty. T. J.. Potter, W. R., Boyle, D. G., Somer, S., and Moan, J.Photobleaching of porphyrins in photodynamic therapy and implications for therapy.Photochem. Photobiol.. 45: 501-506. 1987.

67. Roberts, W. G.. Smith, K. M., McCullough, J. L., and Berns, M. W. Skin photosen-

sitivity and photodestruclion of several potential photodynamic sensitizers. Photochem. Photobiol.. 4V: 431-»38, 1989.

2580

Research. on August 20, 2021. © 1993 American Association for Cancercancerres.aacrjournals.org Downloaded from

1993;53:2571-2580. Cancer Res   Louis Cincotta, James W. Foley and Anthony H. Cincotta  Benzophenoxazine Analogues in EMT-6 Murine Sarcoma CellsPhototoxicity, Redox Behavior, and Pharmacokinetics of

  Updated version

  http://cancerres.aacrjournals.org/content/53/11/2571

Access the most recent version of this article at:

   

   

   

  E-mail alerts related to this article or journal.Sign up to receive free email-alerts

  Subscriptions

Reprints and

  .pubs@aacr.orgDepartment at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

  Permissions

  Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/53/11/2571To request permission to re-use all or part of this article, use this link

Research. on August 20, 2021. © 1993 American Association for Cancercancerres.aacrjournals.org Downloaded from