Photodynamic Effect of a New Photosensitizer ATX-S10 on Corneal Neovascularization

Exp. Eye Res. (1998) 67, 313–322 Article Number : ey980527

Photodynamic Effect of a New Photosensitizer ATX-S10 on Corneal Neovascularization Y U K O G O H TOa, *, A K I R A O B A N Aa, K E N J I K A N E D Ab    T O K U H I K O M I K Ia a

Department of Ophthalmology and b Department of Anatomy, Osaka City University Medical School, Osaka, Japan (Received Lund 7 January 1998 and accepted in revised form 27 April 1998) In order to elucidate the mechanism by which a new photosensitizer ATX-S10 causes the photodynamic effect on neovasculature, we investigated the kinetics and localization of dye accumulation in the neovascular cornea of rats after systemic administration and the development of vascular injury induced by subsequent laser irradiation, compared to those in the normal iris. Under a fluorescence microscope, the neovascular cornea always exhibited more intense fluorescence than the iris between 0±5 and 4 hr after ATX-S10 administration, indicating the preferential deposit of dye in the former tissue. The fluorescence was found inside the vascular lumen at the earliest time period and thereafter in the vascular lining cells, interstitial tissue and infiltrating neutrophils until 6 hr. As observed using light and electron microscopy, laser irradiation performed 2±5 hr after ATX-S10 injection caused extensive vascular thrombosis with endothelial destruction, which persisted for at least 3 days. The proportion of thrombosed vessels at 6 hr after laser irradiation in the neovascular cornea (64³5 % ; n ¯ 3) was significantly (P ! 0±01) higher than that in the normal iris (44³8 % ; n ¯ 3). In the non-thrombosed vessels from heparinized rats, in which thrombosis-related ischemic effect was excluded, mitochondrial vacuolation was the pathologic change commonly seen in the endothelial cells, pericytes and neutrophils. Morphometric analysis revealed that the mitochondria of endothelial cells in the corneal new vessels were more severely injured than those in the iris vessels. The present results indicate that ATX-S10 is a potent photosensitizer which induces photodynamic occlusion particularly of new vessels probably due to the preferential biodistribution of dye in the neovascular tissue. # 1998 Academic Press Key words : ATX-S10 ; corneal new vessels ; fluorescence microscopy ; photodynamic therapy ; rat.

1. Introduction Photodynamic therapy (PDT) was originally developed by Dougherty and colleagues (Dougherty et al., 1978 ; Dougherty, 1984) as an experimental modality for the treatment of solid tumors. Patients were treated with various kinds of malignant tumors including breast and colon cancers by using hematoporphyrin derivatives (HPDs) as photosensitizing agents. Although the mechanism by which PDT causes necrosis of tumor is not fully understood, it is assumed that impairment of tumor microvessels rather than direct toxicity against tumor cells may be responsible for this event (Nelson et al., 1988). In the field of ophthalmology as well, the occlusive effect of PDT on ocular vessels has been obtained through thrombus formation (Royster et al., 1988 ; Kliman et al., 1994b ; Schmidt-Erfurth et al., 1994 ; Peyman et al., 1997). This finding led to the attempt to apply PDT to the treatment of ocular diseases associated with neovascularization in the cornea (Huang et al., 1989 ; Corrent et al., 1989 ; Pallikaris et al., 1993 ; Tsilimbaris et al., 1994 ; Schmidt-Erfurth et al., 1995 ; Obana et al., 1996 ; Soliman et al., 1997), iris (Packer et al., 1984 ; Miller et al., 1991), subretina (Thomas and Langhofer, * Correspondence and requests for reprints should be addressed to: Yuko Gohto, Department of Ophthalmology, Osaka City University Medical School, 1–4–3 Asahimachi, Abeno-ku, Osaka 545–8585, Japan.

0014–4835}98}090313­10 $30.00}0

1987), and choroidal tissue (Kliman et al., 1994a ; Miller et al., 1995, Kramer et al., 1996 ; Husain et al., 1996 ; Bauma et al., 1996 ; Mori et al., 1997). Several first-generation photosensitizers including rose bengal (Royster et al., 1988 ; Huang et al., 1989 ; Corrent et al., 1989) and dihematoporphyrin ether} ester (Packer et al., 1984, Obana et al., 1996) were developed, but proved to be inappropriate for the clinical use because of either prolonged cutaneous photosensitivity resulting from slow-rate excretion or low photodynamic activity. Recently, Sakata et al. synthesized a water-soluble photosensitizer ATX-S10 with a longer absorption wavelength, greater absorption efficiency, lower skin photosensitivity, stronger photooxygenation activities and a higher value of LD (Nakajima et al., 1992, 1995) as &! compared to the first-generation photosensitizers. The agents possessing these advantages are termed secondgeneration photosensitizers and include chloro-aluminium sulfonated phthalocyanine (Miller et al., 1991 ; Pallikaris et al., 1993 ; Kliman et al., 1994a ; 1994b ; Tsilimbaris et al., 1994), benzoporphyrin derivative (Schmidt-Erfurth et al., 1995 ; Miller et al., 1995 ; Kramer et al., 1996 ; Husain et al., 1996), tin ethyl etiopurpurin (Bauma et al., 1996 ; Soliman et al., 1997 ; Peyman et al., 1997) and mono--aspartyl chlorin e6 (Mori et al., 1997), all of which are now under investigation for the clinical application. In the present study, in order to demonstrate the # 1998 Academic Press

314

photodynamic potential and selectivity for new vessels of ATX-S10, the kinetics and biodistribution of this agent in the neovascular cornea and development of PDT-induced vascular injury were analysed, comparing to those seen in the normal iris. 2. Materials and Methods Animals Fifty male Wistar rats (250–350 g b.w.) were purchased from Nippon Biosupply Center (Tokyo, Japan) and housed in a pathogen-free environment with free access to a standard chow diet and water. They were divided into seven groups as shown in Table I according to the purposes of experiments. Photosensitizer Fluorescent dye, ATX-S10 [13,17-bis(1-carboxypriopionyl) carbamoylethyl-8-ethenyl-2-hydroxy-3hydroxyiminoethylidene-2,7,12,18-tetramethhyl porphyrin sodium, m.w. ¯ 927±79] was a gift from Tokyo Hakka Kogyo (Okayama, Japan). This substance has a major absorption peak at 401 nm with minor peaks at 510, 557, 611 and 670 nm. It was diluted with distilled water to a concentration of 10 mg ml−" and administered as a bolus injection via the tail vein at a dose of 16 mg kg−" b.w. Measurement of Plasma Levels of ATX-S10 Three untreated rats were used. After the intravenous (i.v.) administration of ATX-S10, 0±2 ml of the peripheral blood was taken by cardial puncture successively at 5, 30 min, 1, 2, 3, 4, 5, 6, 7, 12 and 24 hr. The plasma obtained after centrifugation was diluted to 1 : 50 with saline. The absorbance of the plasma at the wavelength of 401 nm was measured with a spectrophotometer UV-1200 (Shimadzu, Tokyo, Japan), and the plasma concentration of ATXS10 was determined from a standard curve. Experimental Corneal Neovascularization Corneal neovascularization was induced in both eyes of remaining 47 rats by the intracorneal suture technique. Animals were anesthetized by an intraperitoneal injection of 40 mg kg−" b.w. of pentobarbital sodium. Three 8}0 virgin silk intrastromal sutures were placed end-to-end, starting from the 10, 12 and 2 o’clock positions of the corneal limbus toward the center, without burying knots. Dibekacin sulfate solution was applied at the end of the procedure. Seven to ten days after the operation when new vessels reached the end of the sutures, the eyes were used for following analyses. All procedures were conducted in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research.

Y. G O H T O E T A L.

Fluorescence Microscopy Under the anesthesia with pentobarbital sodium, four eyes were enucleated from two rats each at 1±5, 5, 30 min, 1, 2, 4, 6 and 24 hr after ATX-S10 administration. As a control, neovascular eyes without dye administration were used. Eyes were fixed in 4 % paraformaldehyde in 0±1  phosphate buffer, pH 7±4, overnight at 4°C, then embedded in OCT compound (Miles Scientific, IL, U.S.A.) and frozen in liquid nitrogen. Cryosections of 5 µm in thickness were cut by a cryostat (Jung CM3000, Leica, Nussloch, Germany) and observed under a fluorescence microscope (BX50-FLA, Olympus, Tokyo, Japan) illuminated with a 50-W mercury lamp and equipped with a bandpass exciter filter (400–440 nm in wavelength), a dichroic beam splitter filter (455 nm) and a bandpass emission filter (679–690 nm). A cooled CCD camera (C4880-07, Hamamatsu Photonics, Hamamatsu, Japan) was used to digitize the images. During the analysis, the intensity of the excitation beam was constant and the exposure time was fixed at 4 s in order to compare the intensity of fluorescence between the pictures of different samples. Pictures were printed using a UPD8810 printer (Sony Co., Tokyo, Japan). Photodynamic Therapy Two and a half hours after ATX-S10 administration, 24 rats were treated with argon laser of 514 nm wavelength (ARGON 900, Coherent, CA, U.S.A.) coupled with a slit-lamp microscope under the above mentioned anesthesia. The eyes were irradiated by directing laser light perpendicularly to the corneal new vessels at 127 mW cm−# for 5 min, a total dose of 38±1 J cm−#. Because it was difficult to irradiate both eyes simultaneously, to obtain the eyes with short time intervals (5, 15, 25 min) after laser irradiation, the two eyes of each rat were used for the purpose of different post-irradiation time periods or only one eye was examined. The beam spot was 1 cm in diameter. The tissue around the cornea was however not irradiated, as unlike human beings, rats have a cornea which entirely occupies the anterior part of the eye and the conjunctiva located behind the eyelids. Furthermore, the tissue surrounding the cornea, i.e., eyelids and underlying conjunctiva, were sheltered from irradiation by the operator’s fingers. Physiological saline was dropped onto the surface of the cornea during irradiation to prevent desiccation. For controls, animals with corneal new vessels were either irradiated without ATX-S10 administration or not subjected to irradiation following ATX-S10 administration. Treatment with Heparin Among the rats with PDT, 10 rats were subjected to the treatment with heparin for the purpose of

P H O T O D Y N A M I C E F F E C T O F A T X-S10

315

T I Summary of the protocols for the experiments performed in this study Treatments Neovascularization

ATX-S10administration

Measurement of plasma ATX-S10 level ® ­ Fluorescence microscopy* ­ ­ ­ ® Light and electron microscopy† ­ ­

Laser-irradiation

Heparintreatment

®

®

®

®

®

®

­

®

­

®

­

®

­

­

®

®

­

­

­

­

­

®

­

­

­

­

®

­

No. of eyes examined (No. of rats used) [explanation for experiments]

(3) 32 (16) [four eyes from two rats each for 1±5, 5, 30 min, 1, 2, 4, 6, 24 hr after ATX-S10 administration] 6 (3) [control rats without ATX-S10] 21 (14) [three eyes from two to three rats each for 5, 15, 25 min, 1, 6, 24, 72 hr after laser irradiation] 2 (1) [a non-heparinized control rat without ATX-S10] 2 (1) [a non-heparinized control rat without laser irradiation] 15 (10) [three eyes from two to three rats each for 5, 15, 25 min, 1, 6 hr after laser irradiation] 2 (1) [a heparinized control rat without ATX-S10] 2 (1) [a heparinized control rat without laser irradiation]

* For fluorescence microscopy, both eyes were used for each animal. † For light and electron microscopy of laser-irradiated eyes, one or two eyes were used for each animal.

administration or subjected to no irradiation after ATX-S10 administration. Light and Electron Microscopy

F. 1. Time course of the plasma levels of ATX-S10 in rats after a bolus injection of 16 mg kg−" b.w. of dye via the tail vein. The peripheral blood was taken from three rats successively at indicated time points after dye administration. Each value represents the mean³..

observing the direct photodynamic effect of ATX-S10 on the cell organelles of vascular endothelial cells by eliminating the influence of PDT-induced thrombosis. A dose of 500 U of heparin sodium (Novo Nordisk A}S, Bagsvaerd, Denmark) was i.v. administered at 10 min before and then every hour after laser irradiation until 6 hr. For controls, heparinized animals with corneal new vessels were either irradiated without ATX-S10

At 5, 15, 25 min, 1, 6, 24 and 72 hr after the termination of laser irradiation, the chest was opened under ether anesthesia. Eyes were perfused via the left ventricle with saline for 1 min and then with 1±5 % glutaraldehyde in 0±062  cacodylate buffer, pH 7±4, plus 1 % sucrose for 5 min. For each eye, the neovascular area in the anterior segment of the eye was cut into six parts which included the cornea and underlying iris. Specimens were post-fixed in 2 % OsO , % dehydrated in ethanol series and embedded in polybed (Polyscience Inc., Warrington, PA, U.S.A.). Semithin sections were stained with toluidine blue and observed by light microscopy. Thin sections were stained with uranyl acetate and lead citrate and observed under a JEM-1200EX electron microscope (JEOL, Tokyo, Japan) at 80 kV. Quantitative Analysis The numbers of thrombosed and non-thrombosed vessels in the cornea and iris were counted in six toluidine blue-stained sections prepared from six parts of the anterior segment of the eye, and the percentage

316

Y. G O H T O E T A L.

of thrombosed vessels was calculated. The average percentage in six sections was then calculated for each eye. Three eyes obtained from two or three rats were examined at each time point. The frequency of damaged mitochondria in the endothelial cells of corneal new vessels and iris vessels was analysed using electron microscopy 6 hr after PDT in heparinized rats. Three animals were analysed. Data were expressed as mean value³standard deviation (..). Significant differences between the two populations were analysed using unpaired Student’s t-test.

3. Results Time Course of Plasma Levels of ATX-S10 after Systemic Administration Intravenously injected ATX-S10 was rapidly eliminated from the circulation (Fig. 1) ; the percentage of plasma concentration to the value at 5 min decreased to 59±2 % at 30 min, 43±8 % at 1 hr, 32±0 % at 2 hr and less than 3±0 % at 7 hr after dye injection. The half-life in the plasma as calculated from the data from Fig. 1 was approximately 45 min.

Localization of ATX-S10 in Neovascular Cornea and Normal Iris Under a fluorescence microscope, there was bright fluorescence found inside the lumen of new vessels in the cornea at 1±5 min after dye injection [Fig. 2(a)]. Intraluminal fluorescence was, however, rapidly weakened at later than 30 min [Fig. 2(b), (c)]. On the other hand, the fluorescence in the vascular walls and interstitial tissue was not prominent at 1±5 min [Fig. 2(a)], but increased markedly at 5 min (data not shown) and 30 min [Fig. 2(b)], maintaining the brightness until 4 hr, which had almost completely disappeared at 6 h [Fig. 2(c)]. Under higher magnification, fluorescent dye was localized in the vascular linings, interstitial tissue and infiltrating neutrophils [Fig. 3(a)–(d)]. The vascular lining cells were thin and completely surrounded the lumen, which indicates that they are in fact endothelial cells. The neutrophils as identified by H-E stained serial cryosections (data not shown) were most brightly stained with fluorescent dye [Fig. 3(a)–(d)]. The degree of neutrophil infiltration varied considerably by site in the cornea. In the iris as well, the fluorescence was seen both in the vascular walls and interstitial tissue at 5 min [Fig. 4(a)]. The fluorescence started to decrease as early as 30 min [Fig. 2(b)] and 1 hr [Fig. 4(b)]. It had completely disappeared from both the corneal and iris stroma by 24 hr (data not shown). In the control eyes without ATX-S10 administration, no fluorescence was detected in any tissue components of the cornea and iris (data not shown).

F. 2. Low-power fluorescence microscopy of neovascular cornea (C) and normal iris (I) at 1±5 min (a), 30 min (b) and 6 hr (c) after ATX-S10 injection. Arrows indicate the new vessels in the cornea. (a) Bright fluorescence is seen inside the lumen of blood vessels. (b) The cornea is more intensely stained with fluorescent dye than the iris. Fluorescence is located in the vascular linings and interstitial tissue. Neutrophil infiltration is not prominent in this section. (c) Fluorescence has almost disappeared from both vascular walls and interstitial tissue.¬225.

Histopathology of Vascular Injury after PDT with ATX-S10 At 5 min after PDT, platelets began to adhere to the endothelial cells of corneal new vessels. Thrombi were formed in some vessels between 15 min and 1 hr, and prevailed at 6 hr [Fig. 5(a)]. At 24 hr, the endothelial

P H O T O D Y N A M I C E F F E C T O F A T X-S10

317

F. 4. High-power fluorescence microscopy of the normal iris at 5 min (a) and 1 hr (b) after ATX-S10 injection. (a) Fluorescence is found in both vascular walls and interstitial tissue. (b) Fluorescence is considerably weakened at this time point.¬450.

action of ATX-S10 on new vessels, the percentage of thrombosed vessels was measured in the neovascular cornea and normal iris at various time intervals after PDT (Fig. 6) ; it started to go up at 15 min and reached 63±6³4±5 % (n ¯ 3 eyes) at 6 hr in the cornea, while it remained zero until 1 hr and then increased to 43±6³8±0 % [n ¯ 3 eyes ; significantly (P ! 0±01) low vs. cornea] at 6 hr in the iris. Subcellular Changes in the Vascular Endothelial Cells after PDT with ATX-S10

F. 3. High-power fluorescence microscopy of neovascular cornea at 30 min (a), 1 hr (b), 2 hr (c), and 4 hr (d) after ATX-S10 injection. Fluorescence is seen in the vascular lining cells (arrows) and interstitial tissue rather than inside the lumen of new vessels. The brightest dots seen around the vessels and in the interstitial tissue represent infiltrating leukocytes.¬450.

cells of thrombosed vessels often underwent necrosis, resulting in the disruption of cell organelles and plasma membranes [Fig. 5(c)]. These features representative of the irreversible damage of blood vessels persisted for at least 3 days. In the control eyes treated with either ATX-S10 alone or irradiation alone, the lumen of corneal new vessels were free from erythrocytes, and neither thrombus formation nor endothelial damage was observed (data not shown). In order to demonstrate the selective photodynamic

For the purpose of detecting the cell organelle to which the photodynamic toxicity of intracellular ATXS10 was directed, we eliminated the influence of thrombosis-induced ischemia by repeated administration of heparin. With this procedure, the percentage of thrombosed vessels at 6 hr significantly decreased from 63±6³4±5 % (n ¯ 3 eyes) to 19±9 %³4±0 % (n ¯ 3 eyes) in the cornea and from 43±6³8±0 % (n ¯ 3 eyes) to 0±0³0±0 % (n ¯ 3 eyes) in the iris. The ultrastructural changes of cell organelles were then investigated in the endothelial cells of nonthrombosed vessels. In the neovascular cornea, mitochondrial vacuolation was observed at 30 min after PDT and prevailed at 6 hr [Fig. 7(a), (b)]. This pathologic change included both matrix loss and various degrees of cristae destruction [Fig. 7(c)]. Similar mitochondrial change was found in the pericytes and neutrophils infiltrating the stroma [Fig. 7(a), (b)]. Other cell organelles such as rough endoplasmic reticulum, Golgi apparatus and lysosomes were not damaged [Fig. 7(b)]. There were no degenerative changes in the basement membrane and stromal collagen fibers [Fig. 7(b)]. In the iris, on the other hand, mitochondrial injury was less extensive ; although the mitochondria matrix was dissolved, cristae were usually intact [Fig. 7(d)]. The severity of mitochondrial damage was classified into three : ‘®’,

318

Y. G O H T O E T A L.

F. 5. (a) Light microscopy of the cornea (C) and iris (I) at 6 hr after PDT. Many corneal new vessels (arrows) are occluded by aggregated erythrocytes even after the sufficient perfusion with saline, while, in the iris, occluded vessels (arrows) are not so frequent as in the cornea and many vessels (arrowheads) remain open. Toluidine blue stain.¬210. (b, c) Electron microscopy of the corneal new vessels at 6 hr (b) and 24 hr (c) after PDT. (b) The vascular lumen is occluded by erythrocytes (R) and platelets (arrows) which stick to the endothelial cells (E). P, pericyte. (c) The endothelial cell is entirely disrupted, leaving the remnants of cell organelles (asterisks). The basement membrane (b), however, appears intact. (b)¬6100 ; (c)¬12 000.

almost no degeneration ; ‘­’, loss of mitochondrial matrix ; and ‘ ­­ ’, destruction of mitochondrial cristae. Quantitative analysis demonstrated that mitochondrial injury was more severe in the corneal new vessels than was in the iris vessels (Table II). In the control eyes treated with either laser irradiation or ATX-S10 injection, no ultrastructural changes were seen in the vascular endothelial cells.

4. Discussion

F. 6. Time course of the percentages of thrombosed vessels in the neovascular cornea (*) and iris (E) after PDT. Thrombosed vessels were identified in toluidine blue-stained sections. Three eyes from two or three rats were used for each time point. Each value represents the mean³.. * P ! 0±01 vs. iris vessels.

The present study demonstrated that a newly developed photosensitizer ATX-S10 gave the potent photodynamic effect on the corneal neovasculature through thrombus formation and endothelial destruction, which persisted for at least 3 days, suggesting the irreversible occlusion of the vessels. Furthermore, ATX-S10 was rapidly eliminated from the circulation

P H O T O D Y N A M I C E F F E C T O F A T X-S10

319

F. 7. Electron microscopy of the corneal new vessels (a–c) and normal iris vessels (d) at 6 hr after PDT in heparinized rats. (a, b) Endothelial cells (E) and pericytes (P) exhibited mitochondrial vacuolation (arrows). Neutrophils (N) infiltrating in the interstitial tissue also show the vacuolation of mitochondria (arrows). The Golgi apparatuses (g) are well preserved. (c) Highpower electron microscopy of mitochondrial injury in the endothelial cell of corneal new vessels. Mitochondrial cristae and matrix have disappeared, resulting in vacuolar appearance (arrows), while other cell organelles such as rough endoplasmic reticulum (r) and lysosomes (l) are not damaged. (d) High-power electron microscopy of the endothelial cell of iris vessels shows less severely injured mitochondria (arrows) compared to those in the corneal endothelial cells. Although mitochondrial matrix is lost to some extent, the cristae remain intact. (a)¬9600 ; (b)¬15 000 ; (c)¬25 000 ; (d)¬33 000.

and the ocular tissue. From these properties, this photosensitizer is considered to be promising for therapeutic use in human beings. Another advantage of this agent is the relatively

high selectivity in the photodynamic action on new vessels. This selectivity is considered to be attributable to the preferential accumulation of dye in the neovascular tissue as demonstrated by fluorescence

320

Y. G O H T O E T A L.

T II The degree of mitochondrial injury in the endothelial cells of the corneal new vessels and iris normal vessels at 6 hr after PDT† Degree of mitochondrial injury‡ ® ­ ­­ Normal vessels in 74±0³9±2 the iris New vessels in 22±0³7±0* the cornea

22±7³9±2

3±3³3±1

28±7³7±0

49±3³4±2*

* P ! 0±01 vs normal vessels. † Non-thrombosed vessels from heparinized rats were examined. Fifty mitochondria were analysed per eye by electron microscopy. Three eyes from two or three rats were examined in each group, and the data was expressed as mean³.. ‡ Degree of mitochondrial injury : ‘®’, almost no degeneration ; ‘­’, loss of mitochondrial matrix ; ‘­­’, destruction of mitochondrial cristae.

microscopy ; the fluorescence in the neovascular cornea was brighter and remained longer than that in the iris. It is reported that the liposomal benzoporphyrin derivative shows similar biodistribution between the neovascular cornea and iris as shown here for ATX-S10, but the detailed localization in the tissue was not described (Schmidt-Erfurth et al., 1995). In the present study, ATX-S10 was localized in the vascular linings. Since the lining cells had an attenuated shape and completely surrounded the lumen, they were considered to be endothelial cells (or endothelial cell-pericyte complex). They exhibited more intense fluorescence compared to the lining cells in the iris, suggesting that ATX-S10 might be more actively taken up by the neovascular endothelial cells. In the case of HPDs, they are bound to plasma low density lipoprotein or transferrin and then incorporated by the endothelial cells through receptormediated endocytosis (Jori and Reddi, 1993). The preferential retention of HPDs in the tumor has thus been explained by this selective affinity for the tumorous neovasculature which highly expresses LDL or transferrin receptors rather than by the leakage characteristic of neovasculature (Roberts and Hassan, 1992 ; Thorstensen and Romslo, 1993). In this study, besides the vascular lining cells, a large amount of ATX-S10 was localized in the interstitial tissue presumably due to the enhanced extravascular leakage of plasma ATX-S10 in the neovascular tissue. The fluorescence dye in the interstitial tissue is considered to contribute to the induction of photodynamic damage of new vessels. It was difficult to identify by fluorescence microscopy the cell organelles in which ATX-S10 was accumulated because the fluorescence was equally distributed in the cytoplasm without any appreciable deposits. Electron microscopy, however, demonstrated a degenerative change restricted to mitochondria, i.e.,

mitochondrial vacuolation. This event was commonly seen in the endothelial cells, pericytes and infiltrating neutrophils, all of which showed positive stain with dye by fluorescence microscopy. Furthermore, it was more severe in the corneal new vessels compared to that in the iris vessels, consistent with the present finding on the biodistribution of fluorescent dye between them. The mitochondrial injury may be derived from the photodynamic effect of the ATX-S10 localized in or around mitochondria as interpreted from the previous report that mitochondrial damage was induced in vitro by the photodynamic action of dye which was accumulated in the mitochondria (Roberts, Liaw and Berns, 1989). It has been demonstrated that the intracellular localization of hydrophilic photosensitizers varies among agents ; mono-aspartyl chlorin e6 and chloro-aluminum sulfonated phthalocyanine are accumulated in lysosomes (Roberts and Berns, 1989), while cationic dyes (Grossweiner, 1994) and doxycycline (Shea et al., 1988) are in mitochondria. Thrombosis also contributes substantially to the development of PDT-induced vascular injury as reported by many researchers. It has been postulated that light activation of photosensitizers leads to the production of oxygen free radicals, which are toxic to vascular endothelial cells (Ben-Hur and Orenstein, 1991), being followed by the production of procoagulant factors from injured endothelial cells through lipoxygenase and cycloxygenase pathways (Finger, Wieman and Doak, 1991). In this study, PDT with ATX-S10 induced thrombus formation and subsequently led to the irreversible destruction of endothelial cells. The platelet sticking to the vascular endothelial cells was an early event seen after PDT. This phenomenon is considered to indicate the alteration of endothelial surface by the photodynamic action of intraluminal ATX-S10 as superoxide anions can only work within a distance as short as 0±1 µm from the generation site (Moan and Berg, 1992). However, the difference in the frequency of thrombosed vessels between the corneal new vessels and iris vessels is not likely to be explained by the action of intraluminal ATX-S10 alone as convinced by the fact that the plasma concentration of ATX-S10 must be the same between these two vessels. The ATXS10 localized in all of the intracellular, interstitial and intraluminal compartments is thus considered to participate in the induction of photodynamic vascular endothelial injury and occlusion.

Acknowledgements The authors express their sincere thanks to Dr Isao Sakata, Toyo Hakka Co., Okayama, Japan, Mr. Toru Hirano, Hamamatsu Photonics, Hamamatsu, Japan, and Dr Susumu Nakajima, Division of Surgical Operation, Asahikawa Medical College, Asahikawa, Japan, for their valuable advice and encouragements.

P H O T O D Y N A M I C E F F E C T O F A T X-S10

References Bauma, C. R., Puliafito, C. A., Pieroth, L., Schuman, J. S., Lahav, M., Pedut-Kloizman, T., Kiernan, C., Coker, J. G., Wilkins, J. R., Fujimoto, J. G. and Crean, D. H. (1996). Photodynamic therapy (PDT) of experimental choroidal neovascularization with tin ethyl etiopurprin. ARVO Abstracts. Invest. Ophthalmol. Vis. Sci. 37, S122. Ben-Hur, E. and Orenstein, A. (1991). The endothelium and red blood cells as potential targets in PDT-induced vascular stasis. Int. J. Radiat. Biol. 60 (1}2), 293–301. Corrent, G., Roussel, T. J., Tseng, S. C. G. and Watson, B. D. (1989). Promotion of graft survival by photothrombotic occlusion of corneal neovascularization. Arch. Ophthalmol. 107, 1501–6. Dougherty T. J. (1984). Photodynamic therapy (PDT) of malignant tumors. CRC Crit. Rev. Biochem. 2, 83–116. Dougherty, T. J., Kaufman, J. E., Goldfarb, A., Weishaupt, K. R., Boyle, D. and Mittleman, A. (1978). Photoradiation therapy for the treatment of malignant tumors. Cancer Res. 38, 2628–35. Finger, V. H., Wieman, T. J. and Doak, K. W. (1991). Mechanistic studies of PDT-induced vascular damage : evidence that eicosanoids mediate this process. Int. J. Radiat. Biol. 60 (1}2), 303–9. Grossweiner, L. I. (1994). The science of phototherapy. Pp. 139–55. CRC Press : Boca Raton, U.S.A. Huang, A. J. N., Watson, B. D., Hernandez, Z. E. and Tseng, S. C. G. (1989). Photothrombosis of corneal neovascularization by intravenous rose bengal and argon laser irradiation. Arch. Ophthalmol. 106, 680–5. Husain, D., Miller, J. W., Michaud, N., Connolly, E., Flotte, T. J. and Gragoudas, E. S. (1996). Intravenous infusion of liposomal benzoporphyrin derivative for photodynamic therapy of experimental choroidal neovascularization. Arch. Ophthalmol. 114, 978–85. Jori, G. and Reddi, E. (1993). The role of lipoproteins in the delivery of tumor-targeting photosensitizers. Int. J. Biochem. 25 (10), 1369–75. Kliman, G. H., Puliafito, C. A., Stern, D., Borirakchanyavat, S. and Gregory, W. A. (1994a). Phthalocynanine photodynamic therapy : new strategy for closure of choroidal neovascularization. Laser Surg. Med. 15, 2–10. Kliman, G. H., Puliafito, C. A., Grossman, G. A. and Gregory, W. A. (1994b). Retinal and choroidal vessel closure using phthalocyanine photodynamic therapy. Laser Surg. Med. 15, 11–18. Kramer, M., Miller, J. W., Michaud, N., Moulton, R. S., Hasan, T., Flotte, T. J. and Gragoudas, E. S. (1996). Liposomal benzoporphyrin derivative verteporfin photodynamic therapy. Ophthalmology 103, 427–38. Miller, J. W., Stinson, W. G., Gregory, W. A., El-Koumy, H. A., Puliafito, C. A. (1991). Phthalocyanine photodynamic therapy of experimental iris neovascularization. Ophthalmology 98, 1711–19. Miller, J. W., Walsh, A. W., Kramer, M., Hasan, T., Michaud, N., Flotte, T. J., Haimovici, R. and Gragoudas, E. S. (1995). Photodynamic therapy of experimental choroidal neovascularization using lipoproteindelivered benzoporphyrin. Arch. Ophthalmol. 113, 810–18. Moan, J. and Berg, K. (1992). Photochemotherapy of cancer : experimental research. Photochem. Photobiol. 55, 931–48. Mori, K., Katagiri, T., Sano, A., Anzai, K. and Yoneya, S. (1997). Photodynamic therapy for choroidal neovascularization with a hydrophilic sensitizer : NPe6. ARVO Abstracts. Invest. Ophthalmol. Vis. Sci. 38, S93.

321

Nakajima, S., Sakata, I., Takemura, T., Hayashi, H., Kubo, Y., Samejima, N. and Koshimizu, K. (1992). Tumorlocalizing and photosensitization of photochlorin ATXS10. In : Photodynamic therapy and biomedical lasers. (Eds Spinelli, P., Fante Dal, M. and Marchesini, R.). Pp. 531–4. Elsevier Science : Amsterdam. Nakajima, S., Takemura, T. and Sakata, I. (1995). Acoustic, fluorescent diagnosis of malignant lesions using by HAT-D01 and ATX-S10. Soc. Photo-Optical Inst. Eng. 2371, 495–500. Nelson, J. S., Liaw, L. H., Orenstein, A., Roberts, W. G. and Berns, M. W. (1988). Mechanism of tumor destruction following photodynamic therapy with hematoporphyrin derivative, chlorin, and phthalocyanine. J. Natl Cancer. Inst. 80, 1599–605. Obana, A., Gohto, Y., Matsumoto, M. and Miki, T. (1996). Photodynamic therapy of corneal neovascularization using hematoporphyrin derivatives. J. Jpn Soc. Las. Med. 17, 31–43. Packer, A. J., Tse, D. T., Gu, X.-Q. and Hayreh, S. S. (1984). Hematoporphyrin photoradiation therapy for iris neovascularization. A preliminary report. Arch. Ophthalmol. 102, 1193–7. Pallikaris, I. G., Tsilimbaris, M. K., Iliaki, O. E., Naoumidi, I. I., Georgiades, A. and Panagopoulos, I. A. (1993). Effectiveness of corneal neovascularization photothrombosis using phthalocyanine and a diode laser (675 nm). Laser Surg. Med. 13, 197–203. Peyman, G. A., Moshfegi, D. M., Moshfeghi, A., Khoobehi, B., Doiron, D. R., Primbs, G. B. and Crean, D. H. (1997). Photodynamic therapy for choriocapillaris using tin ethyl etiopurpurin (SnET2). Ophthalmic Surg. Lasers 28, 409–17. Roberts, W. G. and Berns, M. W. (1989). In vitro photosensitization I. Cellular uptake and subcellular localization of mono--aspartyl chlorin e6, choloroaluminium sulfonated phthalocyanine, and Photofrin II. Lasers Surg. Med. 9, 90–101. Roberts, W. G., Liaw, L. H. L. and Berns, M. W. (1989). In vitro photosensitization II. An electron microscopy study of cellular destruction with mono--aspartyl chlorin e6, and Photofrin II. Lasers Surg. Med. 9, 102–8. Roberts, W. G. and Hassan, T. (1992). Role of neovasculature and vascular permeability on the tumor retention of photodynamic agents. Cancer Res. 52, 924–30. Royster, A. J., Nanda, S. K., Hatchell, D. L., Tredeman, J. S., Dutton, J. J. and Hatchell, M. C. (1988). Photochemical initiation of thrombosis : Fluorescein angiographic, histologic, and ultrastructural alterations in the choroid, retinal pigment epithelium, and retina. Arch. Ophthalmol. 106, 1608–14. Schmidt-Erfurth, U., Hasan, T., Gragoudas, E., Michaud, N., Flotte, T. J. and Birngruber, R. (1994). Vascular targeting in photodynamic occlusion of subretinal vessels. Ophthalmology 101, 1953–61. Schmidt-Erfurth, U., Hasan, T., Schomacker, K., Flotte, T. J. and Birngruber, R. (1995). In vivo uptake of liposomal benzoporphyrin derivative and photothrombosis in experimental corneal neovascularization. Laser Surg. Med. 17, 178–88. Shea, C. R., Whitaker, D., Murphy, G. and Hasan, T. (1988). Ultrastructural and dynamics of selective mitochondrial injury in carcinoma cells after doxycycline photosensitization in vitro. Am. J. Pathol. 133, 381–8. Soliman, H., Khoobehi, B., Gebhardt, B. M., Kaufman, H. E. and Crean, D. H. (1997). Photodynamic therapy with SNET2 occludes corneal neovascularization. ARVO Abstracts. Invest. Ophthalmol. Vis. Sci. 38, S512. Thomas, E. L. and Langhofer, M. (1987). Closure of experimental subretinal neovascular vessels with DHE

322

augmented argon green laser photocoagulation. Photochem. Photobiol. 46, 881–6. Tsilimbaris, M. K., Pallikaris, I. G., Naoumidi, I. I., Vlahonikolis, I. G., Tsakalof, A. K. and Lydataki, S. E. (1994). Phthalocyanine mediated photodynamic thrombosis of experimental corneal neovascularization :

Y. G O H T O E T A L.

effect of phthalocyanine dose and irradiation onset time on vascular occlusion rate. Laser Surg. Med. 15, 19–31. Thorstensen, K. and Romslo, I. (1993). The transferrin receptor : its diagnostic value and its potential as therapeutic target. Scand. J. Clin. Lab. Invest. 53, 113–20.