Treatment of Corneal Neovascularization with Dietary Isoflavonoids and Flavonoids

Exp. Eye Res. (2000) 71, 483±487 doi:10.1006/exer.2000.0900, available online at http://www.idealibrary.com on

Treatment of Corneal Neovascularization with Dietary Iso¯avonoids and Flavonoids A N TO N I A M . J O U S S E N ab*, K L A U S RO H R S C H N E I D E R c, J UÈ R G E N RE I CH L I N G d, B E RND K IRC HHO F a A N D F R I E D RI C H E . K RU S E c a

Department of Ophthalmology, University of Aachen (RWTH), Aachen, Germany, bSurgical Research Laboratories, Children's Hospital, Boston, MA 02115, U.S.A., cDepartment of Ophthalmology, University of Heidelberg, Heidelberg, Germany and dDepartment of Pharmaceutical Biology, University of Heidelberg, Heidelberg, Germany (Received Lund 16 March 2000, accepted in revised form 14 July 2000 and published electronically 12 September 2000)

The purpose of this study was to investigate the use of dietary iso¯avonoids and ¯avonoids for the treatment of ocular neovascularization. Corneal blood vessels were induced by intrastromal implantation of pellets containing bFGF. Iso¯avonoids and ¯avonoids (Genistein, Fisetin and Luteolin) were dissolved in a microemulsion to increase bioavailability and applied topically in concentrations between 0.5 and 1 ng ml ÿ1. Corneal neovascularization was quanti®ed under the microscope. In comparison to control eyes, all three substances signi®cantly inhibited corneal neovascularization (P 4 0.05). Fisetin had the strongest effect followed by Genistein and Luteolin. No signi®cant topical side effects were observed. We concluded that the iso¯avonoid Genistein and two structurally related ¯avonoids are potent inhibitors of corneal angiogenesis in vivo. The wide distribution of the ¯avonoids in the plant kingdom together with the presented results suggests that ¯avonoids may contribute to the preventive # 2000 Academic Press effect of a plant-based diet on neovascular disease of the eye. Key words: angiogenesis; corneal neovascularization; Genistein; Fisetin; Luteolin.

1. Introduction The formation of new blood vessels is the hallmark of ocular diseases such as proliferative diabetic retinopathy, certain forms of age related macular degeneration or in¯ammatory conditions of the cornea. While certain manifestations of ocular neovascularization are susceptible to laser treatment or surgical intervention, most forms are still untreatable. In the light of the importance of the various forms of ocular neovascular disease, pharmacological intervention or prevention would be bene®cial. Recognition of the potential bene®ts of controlling angiogenesis has led to a search for angiogenesis inhibitors, including the investigation of dietary factors. Much of the research on the etiology of neovascularization, and especially into the role of nutrients in ocular neovascular disease, has focussed on antioxidants such as antioxidant vitamins, carotens or trace minerals (Fekrat and Bressler, 1996). In this context, the intake of green tea has recently been shown to affect corneal neovascularization (Cao and Cao, 1999). Besides green tea, soy products are an integral part of traditional oriental diets. Among various other ingredients, such diets contain ¯avonoids and iso¯avonoids, and these compounds seem to have a * Address correspondence to: Antonia M. Joussen, Surgical Research Laboratories, Children's Hospital, Enders 1025, Harvard Medical School, 300 Longwood Avenue, Boston, MA 02115, U.S.A. E-mail: [email protected]

0014-4835/00/110483‡05 $35.00/0

bene®cial effect on chronic diseases associated with neovascular growth (Kuo, 1997). The initial work of Fotsis and coworkers has shown that ¯avonoids and iso¯avonoids can not only inhibit the proliferation of tumour cells in vitro but can also impair the proliferation and threedimensional orientation of vascular endothelial cells (Fotsis et al., 1997). Combined with the nutritional epidemiological data, these experimental ®ndings suggest that factors from plant-based diets might also be bene®cial for the treatment of neovascular disease. Since the effect of these naturally occurring compounds had not been demonstrated in vivo, we present evidence that local application of Genistein, Fisetin and Luteolin can signi®cantly inhibit the progression of experimentally induced corneal neovascularization in a rabbit model. 2. Materials and Methods Induction of Corneal Vessels by b-FGF Experiments were performed in New Zealand White rabbits (NZW) according to the ARVO resolution on the use of animals and the declaration of Helsinki under observation of German federal laws. All surgical procedures and quanti®cations of blood vessels were performed under intramuscular general anaesthesia [xylazine hydrochloride (5 mg kg ÿ1) and ketamine hydrochloride (35 mg kg ÿ1)]. # 2000 Academic Press

484

A. M. JOUSSEN ET AL.

A corneal micropocket assay was used for the induction of corneal vessels in response to b-FGF administered intrastromally as described before (Joussen et al., 1999). In brief, 50 ml of 2 % methylcellulose (4000 centipoise; Sigma, Deisenhofen, Germany) were loaded with 500 ng human recombinant bFGF (Boehringer Mannheim, Germany) and allowed to dry. Pellets were implanted intrastromally in the 12 o'clock position, 2 mm central to the limbus. Following implantation, the entrance of the runnel was closed with a single 10-0 nylon suture to prevent uncontrolled liberation of the growth factor. Animals were observed daily, and vascular growth was quanti®ed under the operating microscope on day 3, 6, 9, 12 and 16 after surgery. The vessel length and the vascularized area were quanti®ed with a microcaliper as well as on standardized photographs by measuring the vascularized area. Statistical analysis of the data was performed using the two tailed t-test.

Enhancement of the Bioavailability by Microemulsion Due to their chemical structure, the compounds under investigation are insoluble in water. A series of preliminary experiments investigating both rabbits and rats had shown that administration in conventional and experimental ointments such as petroleum jelly did not cause a biological effect (data not shown). In order to increase solubility and to allow penetration into the cornea, we used a microemulsion (Advanced Pharmacological Technology, Klosters, Switzerland). In previous experiments, we had shown that the microemulsion facilitates corneal drug delivery for various molecules, among those even large molecules such as antibodies (Becker et al., 1999). The composition of the microemulsion, an oleo-tensoid, is shown in Table I. The components PEG-20glyerolacetate, glyceride and vitamin E-acetate were mixed at room temperature. Phosphatidylcholine was later added under continuous shaking, which continued until the mixture was clear. The microemulsion was stored at room temperature and protected from light. Genistein (RBI-Sigma, Deisenhofen, Germany), Fisetin and Luteolin (both from Roth, Karlsruhe, Germany) were added to the emulsion. Homogeneous distribution of these substances within the microemulsion was achieved by prolonged stirTABLE I Components of the microemulsion Poly-ethylene-glycol (PEG-)-20-GlycerolOleorizinoleat Vegetable Mono-, Di- and Triglycerides Phosphatidylcholin Vitamin E-acetat

33 % 66 % 0.5 % 0.5 %*

* The microemulsion was developed by R.A. Duss, Adv. Pharm. Technology, Klosters, Switzerland.

ring. From previous cell culture testing, we know that these substances remain active within the microemulsion for up to 8 days. Treatment with Iso¯avonoids and Flavonoids Following earlier experiments evaluating various concentrations of the experimental compounds (data not shown), Genistein and Luteolin were used at a concentration of 0.5 mg ml ÿ1 and Fisetin at 1.0 mg ml ÿ1. Microemulsion alone served as a control. The right eye of each rabbit was treated four times a day by one of the substances whereas the left eye received only the vehicle (also four times). Treatment was initiated on the day of the surgery. Each of the three substances were used on seven animals in three independent experiments. During the follow-up, animals were observed daily under the surgical microscope. 3. Results Hyperaemia of the iris vessels and of the limbal vasculature was observed from day 1 to day 3 after pellet implantation in all animals. Starting on day 3 after surgery neovascular sprouting began at the limbus. Thereafter, control eyes showed a continuous vascular growth towards the pellet. In control eyes, vessel length was 1.1 + 0.2 mm (means + S.D.) on day 6 and increased to 3.9 + 0.7 mm on day 9. In control eyes, newly formed blood vessels reached the pellet at around day 9 and invaded the pellet thereafter. On day 6 after surgery, measurements of the vessel length were almost identical in control eyes and eyes treated with iso¯avonoids and ¯avonoids with a range from 0.9 to 1.1 mm. In contrast, eyes treated with microemulsion containing either Genistein, Fisetin or Luteolin showed signi®cantly shorter vessels on day 9 (P 5 0.05) [Fig. 1(a)]. The mean vessel length was 1.5 + 0.5 mm for eyes treated by Genistein, 2.9 + 0.7 mm for eyes treated by Fisetin, and 2.8 + 0.5 mm for eyes treated by Luteolin [Fig. 1(a)]. Measurements of the vascularized area also showed a signi®cant reduction of neovascular growth by the compounds under investigation. On day 6 the vascularized area in the control eyes was 4.8 + 2.7 mm2. As shown in Fig. 1(b), the vascularized area on day 6 was reduced to 3.0 + 1.0 mm2 in eyes that received Luteolin (P ˆ 0.05), 2.0 + 1.0 mm2 in eyes that received Genistein (P ˆ 0.02) and 2.3 + 0.9 mm2 in eyes that received Fisetin-containing microemulsion (P ˆ 0.02). Similarly, on day 9, treatment with all of the three substances under investigation resulted in a smaller vascularized area than treatment with the vehicle alone [Fig. 1(b)]. On day 9, control eyes showed a vascularized area of 15.2 + 4.3 mm2 whereas the vascularized area in eyes treated with

D I E TA R Y FAC TO R S A N D CO R N E A L A N G I O G E N E S I S

485

Topical treatment with microemulsion alone or iso¯avonoids dissolved in microemulsion did not result in visible corneal epithelial damage when stained by ¯uorescein and rose bengal. None of the eyes in the treatment groups, except for two eyes treated with Luteolin, showed conjunctival hyperemia. No in¯ammatory reaction of the anterior segment was noted under the microscope. The mild conjunctival injection in two eyes treated with Luteolin did not worsen during the course of the treatment and disappeared two days after discontinuation of the treatment and was not associated with increased angiogenesis. 4. Discussion

F IG . 1. (a) Vessel length in mm after surgery. Vessel length in control eyes showed a continuous increase from observation day 6 to day 12. On day 9 of treatment, eyes treated with microemulsion containing either Genistein, Fisetin or Luteolin showed signi®cantly shorter vessels (P ˆ 0.05). All data shown are means + S.D. (b) Vascularized area in mm2. As compared to the vessel length, the vascularized area in the control eyes increases during the observation period. Treatment with Genistein and Fisetin resulted in a signi®cant inhibition of the vascularized area at all observation points. Luteolin showed signi®cant inhibition on day 12 of treatment (P 5 0.05). All data shown are means + S.D.

Luteolin was 12.9 + 3.4 mm2 (P 4 0.05), in eyes treated with Genistein 4.9 + 1.9 mm2 (P ˆ 0.005) and 8.6 + 2.6 mm2 (P ˆ 0.05) in eyes receiving Fisetin-containing microemulsion. The difference became even more obvious on day 12 when all three substances under investigation signi®cantly reduced the vascularized area in comparison to the control (Fig. 2).

Dietary ¯avonoids represent a family of polyphenol compounds, which are commonly found in plants. Two groups have been isolated from human body ¯uids: lignans, which are derived from ¯avonoid precursors, and iso¯avonoids. Both have molecular weights and structures similar to those of steroids (Fig. 3) (Adlercreutz and Mazur, 1997). Flavonoids are isomeric to iso¯avonoids, and are more abundant in nature compared to the latter. We investigated both groups with regard to their anti-angiogenic activity. Soybeans, which are the richest source of iso¯avonoids, contain up to 1000±3000 mg g ÿ1 of the glycosides of the two iso¯avones Diazein and Genistein. Only small amounts of the substances, however, occur in free form (Adlercreutz and Mazur, 1997). Genistein itself is present at a 30-fold higher concentration in urine of individuals consuming a soyrich diet. This has led to the speculation that a soybean-rich diet might inhibit angiogenesis by virtue of its phytochemical components. In support of this hypothesis, recent experimental data from mice indicate that a diet containing soy protein with 1.0 % phytochemical concentrate leads to a 40 % reduction of transplantable human prostate cancer when compared to a casein control diet (Zhou et al., 1999). Furthermore, histological investigation of tumour tissue showed reduced microvessel density indicating an anti-angiogenic effect of the phytochemical components in soybeans. Although great caution has to be exercized when comparing laboratory data originating from animals to the situation in humans, a variety of epidemiological data suggest that patients with chronic diseases associated with neovascular growth might bene®t from a mostly vegetarian diet (Adlercreutz and Mazur, 1997; Fotsis et al., 1997). In addition, the incidence of breast cancer, prostate cancer and endometrial cancer (all of which are dependent on hormones and neoangiogenesis), as well as coronary heart disease, have a lower incidence in Asia than in Western countries (Fotsis et al., 1993). Concerning neovascular disease in ophthalmology, no direct indication of a therapeutic effect of dietary derived iso¯avonoids and

486

A. M. JOUSSEN ET AL.

F IG . 2. Typical appearance of corneas after 12 days of treatment with iso¯avonoids. Neovascularizations reached the implanted pellet in control eyes (A); thinner and shorter vessels were observed after treatment with Genistein (B) and Fisetin (C). The Fisetin treated cornea shows typical vascular loop formation after stagnation of the neovascular growth. Luteolin (D) also reduced the vascularized area compared to control eyes, but exhibited less in¯uence of vascular hyperaemia if compared to the other iso¯avonoids.

¯avonoids has been obtained. However, the last decades have seen a gradual change in the dietary habits in industrialized countries in Asia away from the classic soy-rich diet. It might, therefore, be possible that the steady increase in the number of patients with exudative macular degeneration observed in Japan (Yuzawa et al., 1997), is partly due to a change in the overall intake of phytochemicals. Our ®ndings, as well as an earlier pilot study applying Genistein subconjunctivally, indicate that all three compounds under investigation inhibit angiogenesis in vivo in the rabbit cornea. Our results are in accordance with investigations indicating that iso¯avonoids can inhibit in vitro angiogenesis (Fotsis et al., 1997). The common ®nding in these studies as well as in our in vivo data is that the anti-angiogeneic effect requires supraphysiological drug concentrations. Genistein inhibited in vitro angiogenesis with half maximal effect at 150 mM (Fotsis et al., 1997). If the human diet is supplemented with 2 mg kg ÿ1 iso¯avones in soy milk, that leads to a mean urine excretion of 67 mmol d ÿ1 equal to urine concentrations of 50 mM and plasma concentrations of 5 mM (Zhou et al., 1999). Consumption of 20 g soy protein per day contains about 23 mg Genistein. In

our in vitro experiments, we used a concentration of 0.5 mg ml ÿ1 applied in eye drops four times daily. Further experiments have yet to show whether the systemic administration of iso¯avonoids can modify ocular angiogenesis. Not much is known about the mechanism by which the anti-angiogenic effect of iso¯avonoids is mediated. From a cell biology perspective, these compounds are anti-oxidants with anti-oxidant protein activities in vitro as well as in vivo. For Genistein, a tyrosin kinase inhibitor effect has been shown. Since the high af®nity receptors of several cytokines (Fotsis et al., 1997), such as epidermal growth factor (EGF) or basic ®broblast growth factor (bFGF) and platelet derived growth factor (PDGF) are tyrosin kinases, the antiangiogenic effect and inhibition of endothelial proliferation could be due to an inhibition of receptor mediated signal transduction. Alternatively, a downregulation of the S6 kinase has been described with similar effects on bFGF, which was used in our experiments to initiate vascular growth. Besides their effect on bFGF-mediated angiogenesis, iso¯avonoids can also down-regulate expression on VEGF and TGF-b (transforming growth factor beta) in nude mice xenografts of breast carcinoma cells (Shao et al.,

D I E TA R Y FAC TO R S A N D CO R N E A L A N G I O G E N E S I S

487

Acknowledgements This work was supported by DFG Jo 324/2-1 (A.M.J.) and a grant from the Land Baden-WuÈrttemberg 116/96 (F.E.K.). The authors thank T. Fotsis, PhD, Laboratory of Biological Chemistry, University of Ioannina, Greece and Professor Dr L. Schweigerer, Department of Pediatrics, University of Essen, Germany, for introducing them to the concept of dietary derived anti-angiogenic factors and for helpful discussion. Ariane Schwartz, Boston, helped in revising the manuscript.

References

F IG . 3. Chemical structures of the compounds used.

1998). The anti-invasive effect of iso¯avonoids, which has been observed in tumour studies, might be mediated through transcriptional down-regulation of MMP-9 and up-regulation of TIMP-1 (Shao et al., 1998), and similar mechanisms are involved in vascular sprouting. In mice fed with soy products, higher apoptosis rates within tumours have been found. Certainly the induction of apoptosis, as well as cell cycle arrest that has been shown for Genistein and other iso¯avonoids (Zhou et al., 1999), takes part in the anti-angiogenic activity. Interestingly, during treatment with Fisetin, the vascularized area regressed. Here, similar mechanisms could be involved as well as an inhibitory effect on vessel maturation. Further studies are needed to determine through which of these possible mechanisms ¯avonoids exert their anti-angiogenic effect on corneal neovascularization.

Adlercreutz, H. and Mazur, W. (1997). Phyto-oestrogens and western disease. Ann. Med. 29, 95±129. Becker, M. D., Kruse, F. E., Azzam, L., Nobiling, R., Reichling, J. and Volcker, H. E. (1999). In vivo signi®cance of ICAM-1Ðdependent leukocyte adhesion in early corneal angiogenesis. Invest. Ophthalmol. Vis. Sci. 40(3): 612±8. Cao, Y. and Cao, R. (1999). Angiogenesis inhibited by drinking tea. Nature 398, 381. Fekrat, S. and Bressler, S. (1996). Are antioxidants or supplements protective for age-related macular degeneration?. Curr. Op. Ophthalmol. 7, 65±72. Fotsis, T., Pepper, M., Edlercreutz, H., Fleischmann, G., Hase, T., Montesano, R. and Schweigerer, L. (1993). Genistein, a dietary-derived inhibitor of in vitro angiogenesis. Proc. Natl Acad. Sci. U.S.A. 90, 2690±4. Fotsis, T., Pepper, M. S., Aktas, E., Breit, S., Rasku, S., Adlerkreutz, H., Wahala, K., Montesano, R. and Schweigerer, L. (1997). Flavonoids, dietary-derived inhibitors of cell proliferation and in vitro angiogenesis. Cancer Res. 57, 2916±21. Joussen, A. M., Kruse, F. E., VoÈlcker, H. E. and Kirchhof, B. (1999). Topical application of methotrexate for inhibition of corneal angiogenesis. Graefe's Archive Clin. Exp. Ophthalmol. 237, 920±7. Kuo, S. M. (1997). Dietary ¯avonoid and cancer prevention: evidence and potential mechanism. Crit. Rev. Oncol. 8, 47±69. Shao, Z. M., Wu, J., Shen, Z. Z. and Barsky, S. H. (1998). Genistein exerts multiple suppressive effects on human breast carcinoma cells. Cancer Res. 58, 4851±7. Yuzawa, M., Tamakoshi, A., Kawamura, T., Ohno, Y., Uyma, M. and Honda, T. (1997). Report of the nationwide epidemiological survey on exudative agerelated macular degeneration in Japan. Int. Ophthalmol. 21, 1±3. Zhou, J. R., Gugger, E. T., Tanaka, T., Guo, Y., Blackburn, G. L. and Clinton, S. K. (1999). Soybean phytochemicals inhibit the growth of transplantable human prostate carcinoma and tumor angiogenesis in mice. J. Nutr. 129, 1628±35.