Suppression of Laser-Induced Choroidal Neovascularization By

2410 Biol. Pharm. Bull. 29(12) 2410—2414 (2006) Vol. 29, No. 12

Suppression of Laser-Induced Choroidal Neovascularization by Subconjunctival Injection of 9a-Fluoromedroxyprogesterone Acetate (FMPA), an Anti-angiogenic Agent, in Rats

,a b b b a Natsuko MURATA,* Taketo YAMAJI, Masayuki UCHIDA, Hiroshi TSUBOI, Hiroto SUZUKI, a c d d d Masashi YAMADA, Tsutomu OIKAWA, Junko NOBUHIRO, Tominari CHOSHI, and Satoshi HIBINO a Pharmaceuticals Development Department, Meiji Dairies Corporation; 1–2–10 Shinsuna, Koto-ku, Tokyo 136–8908, Japan: b Division of Research and Development, Meiji Dairies Corporation; 540 Naruda, Odawara, Kanagawa 250–0862, Japan: c Faculty of Health and Social Work, Kanagawa University of Human Services; 1–10–1 Heiseicho, Yokosuka, Kanagawa 238–8522, Japan: and d Faculty of Pharmacy and Pharmaceutical Sciences, Fukuyama University; 1 Sanzo, Gakuen-cho, Fukuyama, Hiroshima 729–0292, Japan. Received July 11, 2006; accepted August 30, 2006; published online September 14, 2006

9a-Fluoromedroxyprogesterone acetate (FMPA) is a synthetic analog of medroxyprogesterone acetate (MPA). FMPA exhibited more potent anti-tumor and anti-angiogenic activities in some assay systems than the parent agent, MPA. Exudative age-related macular degeneration (AMD) is characterized by choroidal neovascularization (CNV). Anecortave acetate, an angiostatic steroid, is clinically efficacious in patients with exudative AMD. Betamethasone is an anti-angiogenic steroid. Therefore, we examined the effects of FMPA, anecortave acetate and betamethasone on laser-induced CNV in rats. Anecortave acetate and betamethasone were included as positive controls. Crypton laser was applied to the fundus in Brown Norway rats. Laser photocoagulations were performed in each eye between the major retinal vessels of the superior retina. Subconjunctival injection of FMPA, anecortave acetate or betamethasone was performed once just after the photocoagulation (on day 0). The incidence of CNV formation was evaluated by fluorescein angiography (FAG) on day 14. On the next day, examination of the retinal function was performed by electro retinogram (ERG). Subconjunctival injection of FMPA at doses of 300, 1000 and 3000 mg/eye dose-dependently inhibited the incidence of CNV formation. Significant differences were observed at doses of 1000 and 3000 mg/eye of FMPA as compared with the control group. Anecor- tave acetate and betamethasone significantly inhibited the incidence of CNV formation. FMPA at the doses used in this study did not affect the retinal function in rats, as determined by ERG. FMPA appeared to be effective in a rat model of CNV, so it was demonstrated that FMPA might be useful in the treatment of AMD. Key words choroidal neovascularization; age-related macular degeneration; anti-angiogenesis agent; 9a-fluoromedroxyprogesterone acetate (FMPA); electro retinogram

Age-related macular degeneration (AMD) is the primary system.22) FMPA inhibited the activity of plasminogen activa- cause of severe vision loss in old age patients in developed tor (PA), an important protease playing an essential role in countries and exudative (wet) AMD is characterized by the process of angiogenesis, in bovine endothelial cells.23) In choroidal neovascularization (CNV).1,2) Surgical therapy or the mouse dorsal air sac method, FMPA inhibited the mouse photocoagulation therapy has been performed to treat sarcoma 180 cell-induced angiogenesis by oral administra- CNV,3—8) but these therapies affect not only the CNV itself tion.23) These results indicate that FMPA may be useful for but also the healthy retina. Therefore drug therapy for CNV the treatment of diseases associated with angiogenesis. How- with minimal damage to the healthy retina is under investig- ever, there has been no study of the effect of FMPA on CNV. ation and many anti-angiogenic agents have been evalu- Therefore, in this study, we examined the effect of FMPA on ated.9—16) Angiogenesis is an important process by which laser-induced CNV in rats, and also the effects of anecortave new blood vessels are formed from preexisting microvessels acetate and betamethasone as positive controls on laser-in- in many diseases such as cancer, diabetic retinopathy, and ex- duced CNV in rats. In addition, we also examined the effect udative AMD.17) It has been reported that anecortave acetate, of FMPA on the retinal function by electro retinogram (ERG) an angiostatic steroid, is clinically efficacious for improving to confirm whether or not FMPA had retinal toxicity. vision in patients with exudative AMD.14) Betamethasone is a corticosteroid with anti-inflammatory and immunosuppres- MATERIALS AND METHODS sive abilities, and it has been reported that betamethasone shows anti-angiogenic activities in some experimental stud- Animals Male 8-week-old Brown Norway (BN) rats ies.18—20) from Charles River Japan were used. These rats were housed We synthesized 9a-fluoromedroxyprogesterone acetate under standard pathogen-free conditions in constant periods (FMPA) in order to explore a more potent anti-angiogenic of 12-h light/dark (23 °C) and were allowed access to food agent than medroxyprogesterone acetate (MPA),21) which has and water ad libitum. All experiments proceeded in accor- been widely used as a therapeutic agent for breast and en- dance with the ARVO Statement for the Use of Animals in dometrium cancers. FMPA showed 100-times greater anti- Ophthalmic and Vision Research. angiogenic activity than MPA in the chorioallantoic mem- Chemicals FMPA was synthesized in our laboratory as brane (CAM) assay system21) and showed a 7-fold stronger reported previously.21) The structure of FMPA is shown in anti-angiogenic effect than MPA in the rabbit corneal assay Fig. 1. Anecortave acetate was synthesized in our laboratory.

Fig. 1. Chemical Structure of FMPA

Fig. 3. Effect of FMPA on the Laser-Induced CNV in Rats The incidences of CNV in FMPA injected groups, anecortave acetate injected group and the betamethasone injected group are shown as percentages of the vehicle injected group. The number in parentheses expresses the number of samples. Each column rep- Fig. 2. Rat Eye after Eight Laser Photocoagulations resents the meanS.E. FAG was performed on day 14. ∗∗ p<0.01, significant difference Fundus photograph shows retina to be edematous immediately after laser photocoag- from vehicle injected group (Dunnett’s multiple comparison test). ## p<0.01, significant ulation on day 0 (arrows) (A). On day 14 after laser photocoagulation, FAG of CNV le- difference from vehicle injected group (Student’s t-test). sion shows fluorescein leakage (arrows) (B). mination, the rats were anesthetized with intramuscular in- Betamethasone was purchased from Wako Pure Chemical In- jection of a 1 ml/kg mixture (7 : 1) of ketamine hydrochloride dustries, Ltd. (Japan). and xylazine hydrochloride, and the pupil was dilated with Evaluation of the Effect of FMPA on Experimental 0.5% tropicamide and 0.5% phenylephrine hydrochloride eye CNV The rats were anesthetized for all procedures with in- drops. The rats under anesthesia were applied with 0.4% tramuscular injection of a 1 ml/kg mixture (7 : 1) of ketamine oxybuprocaine hydrochloride eye drop and an LED electrode hydrochloride (Ketalar; Sankyo, Tokyo, Japan) and xylazine was placed on the eye. An earth electrode was placed at the hydrochloride (Celactal; Bayer, Tokyo, Japan). The pupil was tail and an indifferent electrode was placed in the nose. The dilated with 0.5% tropicamide and 0.5% phenylephrine ERG was recorded in response to a light stimulus (electrolu- hydrochloride (Mydrine P, Santen Pharmaceutical, Osaka, minescence luminance; 3000 cd/m2 (500 mW), time; 150 ms). Japan) eye drops. Crypton laser (MC-7000L, Nidek Co., Ltd., The amplitudes of a-wave and b-wave were measured using Osaka, Japan) was applied to the fundus (spot size 100 mm, SYNAX ER1100 (NEC, Japan) and the ratio of the ampli- power 100 mV, exposure time 0.1 s) in rats. Eight laser photo- tude of b-wave to the amplitude of a-wave was calculated in coagulations were performed in each eye between the major each eye. retinal vessels of the superior retina (day 0). Fundus photo- Statistical Analysis Data are expressed as the mean graph in rat eye after eight laser photocoagulations (day 0) is S.E. Statistical analysis was performed with Dunnett’s multi- shown in Fig. 2A. Four rats in each group were injected sub- ple comparison test or the Student’s t-test. p<0.05 was con- conjunctivally with FMPA (100, 300, 1000, 3000 mg/eye), sidered statistically significant. anecortave acetate (1000 mg/eye), betamethasone (100 mg/ eye) or vehicle (0.4% Tween 80/2.6% glycerol solution, RESULTS 50 ml/eye) once just after the photocoagulation on day 0. On day 14, 10% fluorescein (0.1 ml/rat) was injected intra- Effect of FMPA on Experimental CNV The incidence venously and fundus fluorescein angiography (FAG) was per- of CNV was expressed as percentage of the vehicle injected formed to evaluate CNV formation using the fundus camera group and the incidence of CNV in each group is shown in (Pro III, Kowa Company, Ltd., Japan). Fundus photograph of Fig. 3. In a rat model of CNV, subconjunctival injection of FAG in rat eye on day 14 after eight laser photocoagulations FMPA at doses of 300, 1000 and 3000 mg/eye dose-depend- is shown in Fig. 2B. CNV formation was classified on the ently inhibited the incidence of CNV formation. The inci- basis of a fluorescein angiogram at the laser photocoagula- dences of CNV at doses of 1000 and 3000 mg/eye of FMPA tion spot as follows: negative, no fluorescein leakage; posi- were 62.26.7% and 48.96.5%, respectively, and signifi- tive, fluorescein leakage. The incidence of CNV was calcu- cant differences were observed as compared with the vehicle lated in each eye as the percentage of CNV positive spots to injected group (p<0.01). The incidences of CNV at a dose of total laser photocoagulation spots (8 spots). All rats were 1000 mg/eye of anecortave acetate and at a dose of 100 weighed before the experiment (day 0) and once a week dur- mg/eye of betamethasone were 73.36.2% and 26.77.5%, ing the experiment (day 7, 14). respectively, and significant differences were observed as Evaluation of the Effect of FMPA on Retinal Function compared with the vehicle injected group (p<0.01). There On the day after evaluation of CNV formation by FAG (day was no statistical difference between the incidences of CNV 15), the rats underwent dark adaptation by housing in a dark in FMPA injected group (1000 mg/eye) and that in anecortave room for 1 h. After dark adaptation, under dim red light illu- acetate injected group (1000 mg/eye). 2412 Vol. 29, No. 12

of 3000 mg/eye and betamethasone injected group at a dose of 100 mg/eye, changes of body weight were signiﬁcantly inhibited as compared with the vehicle treated group (p<0.01).

DISCUSSION

AMD is a retinal disease causing severe and irreversible vision loss, and is a major cause of blindness in elderly patients.1,2) AMD is classified into the exudative (wet) form and the atrophic (dry) form; the exudative form destroys vision more quickly than the atrophic form. In the exudative form of AMD, CNV extends through Bruch’s membrane and into the subretinal space and subretinal pigment epithelium. Some therapies such as surgical therapy and photocoagura- tion therapy have been used to treat CNV,3—8) but these therapies affect the normal chorioretinal tissue. Recently, several anti-angiogenic agents, including TNP-470, irsogladine and Fig. 4. Effect of FMPA on Retinal Function in Rats anecortave acetate, have been investigated to verify their po- The ratio of the amplitude of b-wave to the amplitude of a-wave in each group is tential for clinical use as a new drug for CNV.9,11,14,24—26) shown. The number in parentheses expresses the number of samples. Each column represents the meanS.E. Subconjunctival injection of FMPA, anecortave acetate or be- FMPA is a synthetic analog of MPA. It has been widely tamethasone was performed once on day 0. ERG was analyzed on day 15. used as a therapeutic agent for breast and endometrium cancers without unacceptable side effects, and synthesized in order to enhance its anti-tumor and anti-angiogenic activities by fluorinating MPA.21) FMPA has anti-tumor activity against rat mammary cancer and its activity is greater than that of the parent compound, MPA.22) In addition, FMPA has shown greater anti-angiogenic activity than MPA in some assay systems such as the CAM assay system21) and rabbit corneal assay system.22) These findings suggest that FMPA would be- come a potential candidate drug for treatment of CNV. How- ever, there has been no study of the effect of FMPA on CNV. Therefore, in this study, we investigated the effect of FMPA on laser-induced CNV in rats. The laser-induced CNV model in rats has been widely used for study of CNV.10—12,24—27) The essential process of CNV formation in this model, in- Fig. 5. Effect of FMPA on Body Weight in Rats cluding the break-up of the basement membrane, the migra- Changes of body weight in each group are shown. Each point represents the tion and proliferation of vascular endothelial cells and the tu- meanS.E. of 4 rats. Subconjunctival injection of FMPA, anecortave acetate or betamethasone was performed once on day 0. : Vehicle, : FMPA 100 mg/eye, : bular formation, is similar to the naturally-occurring process. FMPA 300 mg/eye, : FMPA 1000 mg/eye, : FMPA 3000 mg/eye, : Anecortave ac- In this study, FMPA inhibited the incidence of CNV forma- etate 1000 mg/eye, : Betamethasone 100 mg/eye. ∗∗ p<0.01, significant difference from vehicle injected group (Dunnett’s multiple comparison test). ## p<0.01, significant tion by subconjunctival injection, and it was thus thought that difference from vehicle injected group (Student’s t-test). FMPA might be useful drug for treating CNV. PA is an important protease playing an essential role in the Effect of FMPA on Retinal Function The ratio of the process of angiogenesis.28) Irsogladine, an anti-gastric ulcer amplitude of b-wave to a-wave in ERG in each group is agent, inhibits the synthesis of PA and angiogenesis,29,30) and shown in Fig. 4. There was no statistical difference between also inhibits the development of experimental laser-induced the ratio of the amplitude of b-wave to a-wave in each FMPA CNV in rats.11) FMPA inhibited the activity of PA in bovine injected group and that in the vehicle injected group. There- endothelial cells and this activity was almost the same as that fore, FMPA did not affect the retinal function at doses of of MPA reported by Ashino-Fuse et al.31) So, one of the 100—3000 mg/eye by ERG in rats. The ratio of the amplitude mechanisms of anti-angiogenic action of FMPA was consid- of b-wave to a-wave in both anecortave acetate injected ered to be the suppression of PA activity. However, FMPA group and betamethasone injected group were not signifi- may have another mechanism different from that of MPA, cantly different from that in the vehicle injected group, thus because FMPA showed greater anti-angiogenic activity than anecortave acetate (1000 mg/eye) and betamethasone (100 MPA. FMPA inhibited the rat mammary tumor-induced mg/eye) did not affect the retinal function. angiogenesis in the rabbit corneal assay22) and the rat mam- Effect of FMPA on Body Weight There were no statisti- mary tumor was induced by 7,12-dimethylbenz[a]anthracene cal differences between the changes of body weight in FMPA (DMBA). It was reported that the DMBA-induced rat mam- injected groups (100, 300, 1000 mg/eye) and that in the vehi- mary tumor cell line produced two angiogenic factors, rAF-1 cle injected group (Fig. 5). There were no statistical differ- and rAF-2, and these two factors did not have a stimulatory ences between the changes of body weight in anecortave ac- effect on vascular endothelial cell proliferation.32) The angio- etate injected group (1000 mg/eye) and that in the vehicle in- genic factors such as vascular endothelial growth factor jected group (Fig. 5). In the FMPA injected group at a dose (VEGF) and fibroblast growth factors (FGFs) have a stimula- December 2006 2413 tory effect on vascular endothelial cell proliferation.33) So, REFERENCES rAF-1 and rAF-2 are considered to be different from VEGF and FGFs. FMPA inhibited the mouse sarcoma 180 cell-in- 1) Seddon J. M., “Retina,” 3rd ed, Vol. 2, ed. by Ogden T. E., Hinton D. duced angiogenesis in the mouse dorsal air sac method23) and R., Schachat A. P., Wilkinson C. P., Ryan S. J. (ed-in-chief), Mosby, St. it was reported that mouse S-180 sarcoma cells rapidly up- Louis, MO, 2001, pp. 1039—1050. regulated endothelial cell expression of basic FGF.34) There- 2) Votruba M., Gregor Z., Eye, 15, 424—429 (2001). 3) Ohji M., Fujikado T., Kusaka S., Hayashi A., Hosohata J., Ikuno Y., fore, it is considered that FMPA may inhibit the angiogenesis Sawa M., Kubota A., Hashida N., Tano Y., Am. J. Ophthalmol., 132, induced by angiogenic factors such as rAF-1, rAF-2 and 888—896 (2001). basic FGF. Further investigations are necessary to elucidate 4) Ichibe M., Imai K., Ohta M., Hasebe H., Yoshizawa T., Abe H., Am. J. the mechanism of anti-angiogenic action of FMPA. Ophthalmol., 132, 164—171 (2001). It is well known that anti-angiogenic steroids inhibit an- 5) Macular Photocoagulation Study Group, Arch. Ophthalmol., 114, 32,35—37) 677—688 (1996). giogenesis. Anecortave acetate exhibited the anti-an- 6) Macular Photocoagulation Study Group, Arch. Ophthalmol., 112, giogenic activity in both a rat model of retinopathy of prema- 480—488 (1994). turity and an intraocular tumor model.38,39) It has also been 7) Macular Photocoagulation Study Group, Arch. Ophthalmol., 111, reported that anecortave acetate is clinically efficacious for 1200—1209 (1993). improving vision in patients with exudative AMD by the pos- 8) Macular Photocoagulation Study Group, Arch. Ophthalmol., 109, 14) 1242—1257 (1991). terior juxtascleral administration (15 mg). It has been re- 9) Ishida K., Yoshimura N., Mandai M., Honda Y., Invest. Ophthalmol. ported that betamethasone shows anti-angiogenic activities in Vis. Sci., 40, 1512—1519 (1999). some experimental studies.18—20) In this study, we evaluated 10) Takehama Y., Kurokawa T., Kitamura T., Tsukahara Y., Akabane S., the effects of anecortave acetate and betamethasone, anti-an- Kitazawa M., Yoshimura N., Invest. Ophthalmol. Vis. Sci., 40, 459— 466 (1999). giogenic steroids, on laser-induced CNV in rats as positive 11) Kuroki A. M., Kitaoka T., Matsuo Y., Amemiya T., Ophthalmic. Res., controls. Anecortave acetate and betamethasone significantly 35, 137—142 (2003). inhibited the incidence of CNV formation as compared with 12) Yanagi Y., Tamaki Y., Obata R., Muranaka K., Homma N., Matsuoka the vehicle injected group at doses of 1000 mg/eye and H., Mano H., Invest. Ophthalmol. Vis. Sci., 43, 3495—3499 (2002). 100 mg/eye, respectively. There was no statistical difference 13) Muranaka K., Yanagi Y., Tamaki Y., Takahashi H., Usui T., Ohashi K., Matsuoka H., Senda T., FEBS Lett., 579, 6084—6088 (2005). between the incidences of CNV in FMPA injected group 14) D’Amico D. J., Goldberg M. F., Hudson H., Jerdan J. A., Krueger D. (1000 mg/eye) and that in anecortave acetate injected group S., Luna S. P., Robertson S. M., Russell S., Singerman L., Slakter J. S., (1000 mg/eye), but the incidences of CNV in FMPA injected Yannuzzi L., Zilliox P., Ophthalmology, 110, 2372—2385 (2003). group (1000 mg/eye) was somewhat lower than that in 15) Kato A., Kimura H., Okabe K., Okabe J., Kunou N., Nozaki M., Ogura anecortave acetate injected group (1000 mg/eye). Therefore, Y., Retina, 25, 503—509 (2005). 16) Bezerra Y., Fuselier J. A., Peyman G. A., Oner H., Drouant G., Coy D. it is suggested that FMPA may be effective for exudative H., Retina, 25, 345—354 (2005). AMD equal to anecortave acetate. While, it is suggested that 17) Battergay E. J., J. Mol. Med., 73, 333—346 (1995). betamethasone might have caused systemic side effects be- 18) Lemus D., Dabancens A., Illanes J., Fuenzalida M., Guerrero A., cause changes of body weight in the betamethasone-injected Lopez C., Biol. Res., 34, 227—236 (2001). group were significantly inhibited compared with the vehi- 19) Illanes J., Dabancens A., Acuna O., Fuenzalida M., Guerrero A., Lopez C., Lemus D., Biol. Res., 35, 339—345 (2002). cle-treated group. In contrast, FMPA showed the inhibitory 20) Zuniga J., Fuenzalida M., Guerrero A., Illanes J., Dabancens A., Diaz effect on changes of body weight only at the highest dose E., Lemus D., Biol. Res., 36, 233—240 (2003). (3000 mg/eye) in this study. Although FMPA showed the in- 21) Sugino E., Fujimori S., Hibino S., Choshi T., Ichihara Y., Sato Y., Ya- hibitory effect on the incidence of CNV formation at doses maji T., Tsuboi H., Murata N., Uchida M., Shimamura M., Oikawa T., Chem. Pharm. Bull., 45, 421—423 (1997). of 300 and 1000 mg/eye, FMPA did not affect the body 22) Uchida M., Tsuboi H., Yamaji T., Murata N., Kohno T., Sugino E., Hi- weight gains at these doses. Therefore, FMPA might be bino S., Shimamura M., Oikawa T., Cancer Lett., 154, 63—69 (2000). effective for the laser-induced CNV model in rats without 23) Yamaji T., Tsuboi H., Murata N., Uchida M., Kohno T., Sugino E., Hi- producing systemic side effects at doses of 300 and 1000 bino S., Shimamura M., Oikawa T., Cancer Lett., 145, 107—114 mg/eye. Further studies are needed to determine the appropri- (1999). 24) Matsuo Y., Li Y., Taniguchi H., Motoda M., Amemiya T., Jpn. J. Oph- ate dose of FMPA. thalmol., 47, 454—458 (2003). In order to confirm whether or not FMPA had retinal toxic- 25) Yanagi Y., Tamaki Y., Inoue Y., Obata R., Muranaka K., Homma N., ity, examination of the retinal function was performed by Invest. Ophthalmol. Vis. Sci., 44, 751—754 (2003). ERG, because the ERG values obtained from BN rats show 26) Bradey M. E., Cheng L., Bartsch D. U., Appelt K., Rodanant N., Berg- no individual difference.40) In this study, FMPA did not affect eron-Lynn G., Freeman W. R., J. Ocul. Pharmacol. Ther., 20, 217— 236 (2004). the retinal function by ERG, so subconjunctival injection of 27) Dobi E. T., Puliafito C. A., Destro M., Arch. Ophthalmol., 107, 264— FMPA seemed to be safe for the treatment of AMD. 269 (1989). In conclusion, it was demonstrated that subconjunctival in- 28) Blei F., Wilson E. L., Mignatti P., Rifkin D. B., J. Cell Physiol., 155, jection of FMPA inhibits the laser-induced CNV formation in 568—578 (1993). rats. It is suggested that FMPA may be useful for exudative 29) Sato Y., Morimoto A., Kiue A., Okamura K., Hamanaka R., Kohno K., Kuwano M., Sakata T., FEBS Lett., 322, 155—158 (1993). AMD. 30) Ono M., Kawahara N., Goto D., Wakabayashi Y., Ushiro S., Yoshida S., Izumi H., Kuwano M., Sato Y., Cancer Res., 56, 1512—1516 Acknowledgment This work was supported in part by a (1996). Grant-in-Aid for Scientific Research on Priority Areas 31) Ashino-Fuse H., Takano Y., Oikawa T., Shimamura M., Iwaguchi T., Int. J. Cancer, 44, 859—864 (1989). 17035068 from The Ministry of Education, Culture, Sports, 32) Oikawa T., Hiragun A., Yoshida Y., Ashino-Fuse H., Tominaga T., Science and Technology (MEXT). Iwaguchi T., Cancer Lett., 43, 85—92 (1988). 33) Folkman J., Shing Y., J. Biol. Chem., 267, 10931—10934 (1992). 2414 Vol. 29, No. 12

34) Peverali F. A., Mandriota S. J., Ciana P., Marelli R., Quax P., Rifkin D. 38) Penn J. S., Rajaratnam V. S., Collier R. J., Clark A. F., Invest. Ophthal- B., Della Valle G., Mignatti P., J. Cell Physiol., 161, 1—14 (1994). mol. Vis. Sci., 42, 283—290 (2001). 35) Crum R., Szabo S., Folkman J., Science, 230, 1375—1378 (1985). 39) Clark A. F., Mellon J., Li X.-Y., Ma D., Leber H., Apte R., Alizadeb 36) McNatt L. G., Weimer L., Yanni J., Clark A. F., J. Ocul. Pharmacol. H., Hegde S., McLenagban A., Maybew E., D’Orazio T. J., Niederkorn Ther., 15, 413—423 (1999). J. Y., Invest. Ophthalmol. Vis. Sci., 40, 2158—2162 (1999). 37) Li W. W., Casey R., Gonzalez E. M., Folkman J., Invest. Ophthalmol. 40) Sakai H., Tani Y., Shirasawa E., Shirao Y., Kawasaki K., Ophthalmic Vis. Sci., 32, 2898—2905 (1991). Res., 27, 57—63 (1995).