0 1992 MUNKSGAARD

Pigment Cell Research 5:379-386 (1992)

The Effect of Oxygen on Melanin Precursors Released from Retinal Pigment Epithelial Cells In Vitro KIYOSHI AKE0,'*273NOR10 UENO,' AND C. KATHLEEN DOREY' 'Department of Ophthalmolo School of Medicine, Keio University, 35 Shinanomachi, Shinjuku-ku, Tokyo 160, JapanFEye Research Institute of Retina Foundation, Boston, MA 02114, U. S.A., 'Department of Ophthalmology, National Defense Medical College, 3-2, Namiki, Tokorozawa-shi, Saitama, 359 Japan. The autoxidation of dopa to melanin in culture media causes toxicity to retinal pigment epithelial (RPE) cells and endothelial cells. The damage is specific to cell type and to the ambient oxygen concentration. 'Ib determine whether RPE cells influence the oxidation of dopa to media, we compared light absorbing dopa derivatives in the media exposed to cells with those found in the media incubated without cells. Dopa was extensively oxidized in the presence of RPE cells, and more light absorbing substances were generated with higher dopa and oxygen concentrations. However, a n increase in ambient oxygen concentration decreased the quantity of several dopa derivatives which had been formed. The data provided evidence that RPE modulated dopa metabolism. Quinolic derivatives produced from a tyrosinase reaction and dopa-melanin formation moved the peak absorbance wavelength of dopa into the visible range. The spectrum between the dopaderived compounds in the media has a n absorbance at 240-275 nm and a maximum around 300 nm wth a shoulder near 375 nm. Gaussian analysis (peak separation) resolved these spectra into five components: a sharp band at 248 nm, a band at 295 nm, a large band at 359 nm, and two broad bands at 459 and 585 nm. Key words: Oxygen, Dopa, Melanogenesis, Spectrum, Pigment epithelium of the eye

INTRODUCTION A series of oxidation starting with the reaction of tyrosinase and dihydroxyphenylalanine (dopa) forms melanin (eumelanin or pheomelanin), a high-molecular-weight complex polymer (Prota, 1980). Wick Byers et al. (19771, F'awelek and Lerner (19781, and Pawelek et al. (1980)reported that some intermediates of melanogenesis were toxic to tissues. Oxygen radicals (singlet oxygen, superoxide anion, hydroxyl radical, and hydrogen peroxide) are important toxic agents in cells (Marx, 1987)and superoxide anions are produced when dopa changes into quinolic derivatives or melanin as described by Hochstein (19831, Tomita et al. (1984), and Sealy e t al. (1984). Dopa methyl ester inhibits DNA synthesis and rapidly arrests the growth of lymphoid cells in the GI-phase (Wick, 1979). Ldopa inhibits the cell cycle of HeLa in the GI-phase and S-phase at high and low doses, respectively. We also observed the same effect of Gdopa on the cell cycle of RPE cells (unpublished) as MM96L (Kable and Parsons, 1989). When dopa was mixed with cell culture media, the toxicity of dopa (unstable quinolic derivatives) to porcine retinal pigment epithelial cells (RPE) was correlated with ambient oxygen partial pressure (Akeo et al. 1989). The inhibitory effects of dopa on the cell cycle vary dependent on

the target cell. For example, vascular endothelial cells (EC) are more sensitive to the melanin-intermediate than R P E and skin fibroblasts(FB) (Akeo e t al. 1989). Because neither EC nor F B could be considered to produce melanin actively and generate its toxic precursors, we speculated that the difference in the sensitivity of these cells to dopa was due to auto-oxidation of dopa. I t is uncertain whether the RPE cells can stimulate the oxidation of dopa in culture media. Because the melanin precursors as well as dopaderivative products are quite unstable, it is extremely difficult to extract them from cell culture media and determine the concentration of these intermediates. To avoid these difficulties, we measured the absorption spectrum of cell culture media and by computer performed the component-analysis of the precursors in the spectrum based on the assumption that these precursors will reach an equilibrium (Maddams, 1980; Ueno and Chakrabarti, 1988).

Address reprint requests to; Kiyoshi Akeo, Department of Ophthalmology, National Defense Medical Collepe, 3-2Namiki, lbkorozawashi, Saitama, 359 Japan.

Received March 4,1992; accepted October 6, 1992.

K. Akeo et al.

380

1.2 1 .o

0.8 0 0

c

+ 0

0.6

0

m 0

Q

0.4

0.2

0.0 200

I

I

I

I

I

I

I

250

300

350

400

450

500

550

600

Wa ve le n g th (nm) Fig. 1. Scanning absorbance of media with dopa in 20% oxygen. No cells: 100 km dopa (open triangles down), 250 pm dopa (open cir-

cles). Culture of RPE cells: 100 Fm dopa (solid triangles down), 250 )*m dopa (solid circles).

HBSS into the subretinal space and removed from the posterior eye cup. The eye cup was solid with trypsin and, after 5 to 10 min, the RPE cells were removed by gentle trituration. These RPE cells were immediately transferred MATERIALS AND METHODS into a centrifuge tube with complete DMEM containing 15% Materials FBS, amphotericin B (2.1 pg/ml), penicillin (83 U/ml), The medium used for cell culture contained the follow- and streptomycin (83 pg/ml). After being washed with the ing: fetal bovine serum (FBS) (Hyclone Laboratories, medium, the cells were collected by centrifugation and plated Logan, UT), Dulbeco's Modified Eagle's Medium (DMEM), in a 35 mm petri dish. When the cells reached confluence, MEM without phenol red (PR-MEM), amphotericin B and after they were trypsinized, we plated them into a 25 (Fungizone, 250 pg/ml), penicillin (10,000U/ml), streptomy- cm2petri dish. These cells, cultured in 95% air/5% C02 in cin (10,000 pg/ml), trypsin (2.5%), Hank's Basal Salt Solu- humidified incubators, were fed complete DMEM at weekly tion (HBSS) (Gibco Laboratories, Grand Island, NY) intervals until confluence. HEPES buffer, and 3,4-dihydroxyphenylalanine(L-dopa) RPE Cell Experiments (Sigma Chemical Co., St. Louis, MO). A 25 em2flask (Falcon, Approximately the same number of three-passage RPE Cockeysville, MD) was used throughout this study. An cells were plated in triplicate sets of 25 cm2 flasks. These oxygen-regulated incubator (VWR 1750, VWR Scientific, cells reached confluence in 1 week under standard condiSan F'rancisco, CA) and a Coulter Counter (Coulter Election (95% air/5% COz) with complete DMEM. Then the tronics, Hialeah, FL) were used for incubation and cellflasks (total: 27) were divided into 9 groups (3flasks each), counting, respectively. and incubated with 0, 100, or 250 pM of dopa with 20% (control), lo%, and 5%oxygen by adding nitrogen with pheCulture of RPE Cells nol red free Eagle's Minimum Essential Medium (PRFresh porcine eyes with excess tissue removed by scis- MEM) containing 20 mM HEPES buffer without serum sors and forceps were dissected under sterile conditions. (S-media) (Thble 1). Because of phenol red functions as an The retina was carefully floated of the RPE by pipetting antioxidant, PR-MEM was used for these experiments.

In this study, we investigated the effect of ambient oxygen on the melanin precursors in cell culture media of porcine RPE by using spectroscopy and component-analysis.

Melanin precursors released from RPE in vitro

381

1.2

1 .o

0.8 a) v

S 0

+

0.6

0 v,

n

Q

0.4

0.2

0.0

200

250

300

350

400

450

500

550

600

Wavelength (nm) Fig. 2. Scanning absorbance of S-media with 250 pm dopa in lowered oxygen concentrations. No cells 5% (open squares), 10% (open tri-

Spectral Analysis of the Media After a specified incubation time (one day) under various conditions, the culture media were removed from the flask and transferred into a centrifuge tube to form a clear supernatant. After pouring the clear medium into a 1 cm cuvette, the absorption spectrum was measured from 200 to 600 nm at a scan speed of 750 nm per min by a Beckman DU-40 spectrometer (Beckmann, Fullerton, CA).A series of data was simultaneously corrected for the baseline, using S-media for the experiments of dopa and oxygen, and incubated in 20% oxygen without dopa and cells for one day to be used as the blank. The spectrum of the media incubated without cells was subtracted from that of identical media incubated with cells. If each component of the melanin precursor compounds has a characteristic spectrum, we need to separate the spectrum and eliminate interaction of adjacent spectra. By using a computer simulation, it is possible to determine which individual spectrum of the compounds is most effected by the oxidation. Gaussian analysis (peak separation) of the absorption spectrum was performed by using spreadsheet software (Quattro Pro, Borland International Inc., Scotts Valey, CA) and a personal computer. The spectrum was resolved into several component bands (Maddams, 1980; Ueno and Chakrabarti, 1988):

angles up), and 20% (open circles) oxygen. Culture of RPE cells: 5% (solid squares), 10%(solidtriangles up), and 20%(solid circles) oxygen. 4

f(x) = C H.exp[-4 ln2((x - XC)/W)~] n=l

where f (x) is the sum of the normal distribution spectra; H, peak height; x, wavelength; xc, wavelength of peak; w, width of the half height of each spectrum. Peak height and width of half height were selected to minimize the differences between the simulated spectrum and the actual data by checking the shape of the graph calculated by the computer.

Statistical Analysis Regression analysis is commonly used to forecast values for a given variable (called the dependent variable) based upon the values of other variables (the independent variables). We wanted to estimate the relative contributions of dopa, oxygen, and the existence of RPE cells (the independent variables) to peak absorbance (the dependent variable). We performed a stepwise linear regression analysis on the data sets under the absence and the presence of RPE cells individually in order to compare the relative sensitivity of the melanin precursor spectra to dopa and oxygen and specify the criteria that select and remove variables. The model used for analysis was:

K. Akeo et al.

382

variable, (IpE)(Dopa),to permit identification of significant RPE cell-specificdifferencesin response to dopa. These analyses generated the values and s i d c a n c e of Pothrough ps. The suitability of the model for these data was evident for each step to the termination of the regression (5; level of significance of the composite model), in view of the fact that the minium F value for a variable continues in regression.

1 dedium without R P E

1.0.

0.8 . a, V

c

RESULTS Spectrum of Dopa-Derived Compounds A typical spectrum (225-600 nm) of S-media in 20% oxygen with 250 pm dopa shows a small peak near 245 nm and a large peak near 300 nm with large shoulder around 375 nm. The addition of dopa to S-media (100, 250 pm) with and without RPE cells increase the media absorbance proportionally. The increase in absorbance by adding dopa to media with RPE cells is larger than that of the control. RPE cells distinctly stimulated the production of melanin derivatives. The increase in dopa concentration from 100 to 250 pm shifted the 300-nm peak to 320 nm (red shift) (Fig. 1).

0

0.6

.

0 v)

n

6

0.4

0.2

0.0 240-250 295-305

375

240-250 295-305

375

Wavelength (nm)

-20%

Oxygen 10% Oxygen 5% Oxygen

Fig. 3. The effects of oxygen on the peak absorption of S-media with 250 p n dopa. 20%(solid), 10% (cross hatched), and 5% (horizontal line) oxygen.

Absorbancei = Po + pl(dopaIi + Pz(oxygen)i + P&qa)(owgen)i + PdIPdi + Ps(dopa)(Ipdi + Pdoxygen)(IpE)i + Ei where the absorbance in an individual plate (i) is considered as a possible function of the plate, the amount of dopa presen in i, the ambient oxygen concentration, a product of interaction of dopa and oxygen [(dopa)(oxygen)],and a random error term (Ei). IpE is an indicator for the existence of RPE cells (equal to 0 or 1 for the absence or presence of RPE cells, respectively). All data from the experiments were analyzed simultaneously using computer programs (MICROSTAT 11, Ecosoft Inc., Indianapolis, IN). We used another

Effects of Dopa-Oxidation on the Spectrum of Dopa-Derived Compounds Figure 2 shows the effect of oxygen concentration on the absorption spectra of S-media that contain RPE cells with 250 pm dopa. Hypoxia suppressed the absorbance with a maximum a t 300 nm and a shoulder near 375 nm independently of the existence of RPE cells. However, lowering the oxygen concentrationsincreased a band spectrum at 240-275 nm in the S-media after the culture of RPE cells but reduced the band spectrum without RPE cells (Fig. 2). The relation between oxygen concentration and peak absorption of S-media with 250 pm dopa is presented in Figure 3. An increase in oxygen concentration enhanced the absorption spectra of S-media with and without RPE cells, except for the peak of media with RPE cells around 245 nm that was inhibited in inverse proportion to the oxygen concentration. 'kble 1 exhibits the results of stepwise linear regression analysis of the data for S-media that contain dopa (0, 100, 250 pm) with varied oxygen concentrations (5, 10, 20%). The positive coefficient values of (Dopa)(IpE)around 245, 300, and 375 nm indicated a significant increase in the melanin precursors due to the presence of RPE. However, the

TABLE 1. Effects of Dopa and Oxygen on Spectrum of Dopa-Derived Compoundsa dopa dopa-oxy dopa-IpE OXY-lPE

B1

B3

Value P

0.0013 0.00001

-

Value

P

0.0012 0.00001

Value P

0.0008 0.00008

Wavelength 240-250 nm 295-305 nm 375 nm

aOxy = oxygen, P = significance. Absorbance = B, + B1 (dopa) + B2 (oxygen) + B3 (dopa) (oxygen) + B4(IPd + B5 (dopa 1 (IPE)+ B6 (oxygen) ( IPE)

B5

B6

Fit of the model (n)

3 = 0.83

0.0018 0.00001

-

0.00003 0.03

0.0022 0.00001

-

8 = 0.93

-

0.00001 (54)

0.00005 0.0002

0.0022

-

0.00001

-

0.040 0.03

0.00001 (54)

3 = 0.95 0.00001 (54)

Melanin precursors released from RPE in vitro

383

1.2 4-

I

1 .o

0.8 a,

u

t

W

0.6 0 v)

n

Q

0.4 a

=

A

0.2

' I

A

'

I

v 1

1 b

' I

I

0.0 200

250

300

350

400

450

500

550

600

W a v e l e n g t h (nm) Fig. 4. Simulated spectrum of dopa-derived compounds in media. The spectrum of media with 250 pm dopa culturing RPE cells in 20% oxygen is indicated by a solid line with solid circles. The spectrum has been resolved into five components bands by using Gausian anal-

ysis (solid squares, open squares, solid triangles up. open triangles up, and solid triangles down). The sum of each separated spectrum calculated from the equation (open circles) overlaps on t h e media spectrum.

presence of RPE and the increase in oxygen concentrations are likely to cause significant decrease in the absorbance around 375 nm because of the negative coefficient value of (OXY)(IPE).

because every coefficient of (dopa)(Ip,) was positive, according to stepwise linear regression analysis of the data for simulated spectra of S-media with 0, 100, and 250 pm of dopa and varying oxygen concentrations (5,10,20%) (7hble 2). However, RPE cells diminished oxygen tension due to generation of substances absorbing light around 248,295, 459, and 585 nm because the negative values of (Oxygen) (IPE)were negative.

Simulation of Spectrum of Dopa-Derived Compounds in Media The absorption spectra (225-600 nm) of S-media in 20% oxygen that contain RPE cells with dopa (250 pm) are shown in Figure 4. Gaussian analysis resolved this composite spectrum into five distinct bands: a sharp band around 248 nm, a band around 259 nm, a large band near 359 nm, and two broad bands near 459 and 585 nm. Effects of Dopa-Oxidation on Simulated Spectra of Dopa-Derived Compounds Hyperoxia stimulated the simulated absorption spectrum a t 359 nm of S-media with RPE cells exposed to 250 pm of dopa but decreased the spectra at 248,295, and 459 nm (Fig. 5A). On the contrary, an increase in oxygen enhanced every simulated spectrum of S-media without RPE cells (Fig. 5B). The RPE cells promoted the production of light absorbing substances of dopa-derived compounds in the media,

DISCUSSION The presence of RPE cells increased the dopa-derived compounds of S-media because the coefficient value of (dopa)(IpE)was positive a t 240-250 nm, 295-305 nm, and 375 nm. RPE cells themselves have tyrosinase activities according to earlier reports (Mannagh et al. 1973; Dryja et al. 1978; Basu et al. 1983; Dorey e t al. 1983). An obvious dependence of the simulated composite band around 360 nm on dopa concentrations for RPE cells suggests that RPE cells produced most of the light absorbing substances around 360 nm in the media. Prota (1980) summarized the reaction scheme for the metabolic pathways leading to eumelanin and pheomelanin that was converted from dopachrome and cysteinyldopa, respectively. I t is possi-

K. Akeo et al.

384

0.8 [

0.8

0.6

Q,

V

c 0

'?

0.4

0

T

v,

n Q

0.2

0.2

0.0

nn V."

2476

2953

Wavelength (nm)

=20%

=2O%

Oxygen

10% Oxygen 5% Oxygen

a

3585

459.1

5854

Wavelength (nm)

10% Oxygen 5% Oxygen

b

Fig. 5 . The effects of oxygen on simulated peak absorption of S-mediawith 250 km dopa (a: media after culture of RPE cells, b: media

Oxygen

without RPE cells). 20% (solid), 10% (cross hatched), and 5% (horizontal line) oxygen.

ble that the amount of 5-S-cysteinyldopa increased in the trum and the peak absorbance around 250,300, and 375 nm, media because the dopaquinone converted from dopa com- in proportion to the dopa concentration. These results indibined with cysteine (It0 et al. 19%) in the presence of RPE cated that the cytotoxicity of RPE cells from exposure to cells. Rorsman et al. (1979)showed that 5-S-cysteinyldopa dopa was related to enhancement of a light absorbing subis present in serum itself. Dryja et al. (1977)have reported stance in proportion to the quantity of dopa added. However, that 5-S-cysteinyldopa and related metabolites are formed an increase in oxygen concentrations decreased the absorand secreted by melanocytes of an adult eye in which the bance around 375 nm in the presence of RPE, and diminished the simulated peak of the media around 250,295,460, melanin synthesis rate is negligible. The interaction of dopa and oxygen, i.e. dopa-oxidation, and 580 nm. It is possible that some extra cellular melahad significant effects on absorbance, with a maximum nin precursors are consumed in order to produce melanin around 300 nm and a shoulder near 375 nm. The presence pigments in RPE cells. Previously, Korytowski et al. (1985) of RPE increased all absorbances in the simulated spec- reported that the melanin pigment competed with SOD for TABLE 2. Effects of Dopa and Oxygen on Simulated Spectrum of Dopa-Derived Compounds OXY dopa-oxy dopa-IpE OXY-fPE dopa Wavelength 247.6

2

1.6nm

295.3

2

2.4 nm

358.5

2

3.3nm

459.1

2

3.0nm

585.4

2

8.8nm

Value

P Value P Value P Value P Value

P

See Table 1for abbreviations.

B1

B2

B3

0.0010 0.00001 0.0010 0.00001 0.0004 0.02 0.0005 0.00001

-

-

-

- 0.004

0.04 -

-0.003 0.003

0.00006 0.0002 -

0.00003 0.00001

B5

0.0018 0.00001 0.0013 0.00001 0.0015 0.00001 0.0014 0.00001 0.0007 0.00001

B6

-0.0042 0.08 - 0.0045 0.05 -

-0.0048 - 0.00003 - 0.0024

0.008

Fit of the model (n) r;! = 0.85 0.00001 (54) 1.2 = 0.82 0.00001 (54) 1.2 = 0.94 0.00001 (54) r;! = 0.93 0.00001 (54) r;! = 0.91 0.00001 (54)

Melanin precursors released from RPE in vitro scavenging of superoxide radicals, and Korytowski e t al. (1986) suggested that scavenging of superoxide radicals by melanin is a possible factor in the photo protection afforded by melanin pigments. The spectral absorbance of the dopa-derived compounds of S-media was at 240-275 nm, with a maximum around 300 nm and a shoulder around 375 nm. These compounds of melanin precursors may be the sum of the spectrum of the components. We tried approximate Gaussian analysis of the components of melanin precursors by using a personal computer, and resolved the composite spectrum of melanin precursols into five positive bands in the media without serum (250, 300, 360, 460, and 580 nm). Thompson e t al. (1985) reported the results of investigations by pulse radiolysis of the oxidation of the melanin precursor, dopa, and the most abundant cysteinyldopa isomer. These studies lead to transient absorption spectra. The oxidation of dopa formed dopasemiquinone with a peak absorbance at 350 nm, which decayed into dopaquinone with an absorption maximum around 380 nm. Melanin precursors produced during dopamelanin formation have not been identified completely due to instability of quinolic derivatives and intermediate products. It is uncertain whether each simulated peak of absorption spectra of media is consistent with the spectrum obtained by pulse radiolysis. Dorey et al. reported that reversed-phase high-performance liquid chromatography (HPLC) on RPE cells and their media detected melanogens, i.e., intermediates of melanin biosynthesis, including several indole derivatives. HPLC analysis of media with Gdopa is expected to yield information about effects of RPE cells and oxidation on melanin precursors in the media. From the result of simulation, however, we are able to conclude that RPE cells increase the quantified substances with the spectrum around 360 nm, and that the melanin precursors decrease in inverse proportion to the oxygen concentration in the presence of RPE cells.

CONCLUSIONS The presence of RPE cells stimulated the oxidation of dopa, while dependent on the concentration of the latter, and the peak absorbance wavelength of dopa shifted to the visible range due to quinolic derivatives produced from tyrosinase reaction and dopa-melanin formation. However, the presence of RPE cells diminished the quantity of some dopa derivatives. These data provided evidence that RPE modulated dopa metabolism. The absorbance spectral range of the dopa-derived compounds in media without serum was 240-275 nm, with a maximum around 300 nm and with a shoulder around 375 nm. Computer simulations of the individual components present in the media resolved into five positive bands. ACKNOWLEDGMENT The authors are grateful to Professor Shigekuni Okisaka (National Defense Medical College) and Professor Yoshihisa Oguchi (Keio University) for their enthusiastic encouragement. This project was supported by a Grant for Scientific Research (No. 02454409) from the Ministry of Education, Japan, Morganstern Scientist Fund, and the E R I Macu-

385

lar Disease Research Center. A part of this paper was presented a t the Annual Spring Meeting of the Association for Research in Vision and Ophthalmology, Sarasota, May 1988, at the 7th Japanese Chapter of the International Society for Eye Research, Osaka, December 1989, and a t the 94th Congress of the Japanese Ophthalmologic Society, Kyoto, May 1990.

REFERENCES Akeo, K., S.A. Curran, and C.K. Dorey (1988) Superoxide dismutase activity and growth of retinal pigment epithelial cells are suppressed by 20% oxygen in vitro. Curr. Eye Res., 7961-967. Akeo, K., D.B. Ebenstein, and C.K. Dorey (1989) Dopa and oxygen inhibit proliferation of retinal pigment epithelial cells, fibroblasts and endothelial cells in vitro. Exp. Eye Res., 49:335-346. Basu, P.K., P. Sarkar,I. Menon, F. Carre, and S. Persad (1983)Bovine retinal pigment epithelial cells cultured in vitro: Growth characteristics, morphology, chromosomes, phagocytosis ability, tyrosinase activity and effect of freezing. Exp. Eye Res., 36:671-683. Dorey, C.K., C. Bekhor, N. Sorgente, and S.J. Ryan (1983) Porcine R P E cells maintain differentiated characteristics in vitro. Ivest. Ophthalmol. Vis. Sci., 24 (Supp1.);142. Dorey, C.K., X. Torres, and T. Swart (1990) Evidence of melanogenesis in porcine retinal pigment epithelial cells in vitro. Exp. Eye Res., 501-10. Dryja, T.P., D.M. Albert, H. Rorsman, E . Rosengren, and T.W. Reid (1977) Presence of cysteinyldopa in the mature bovine eye. Exp. Eye Res., 25:459-461. Dryja, T.P., M. O”ei1-Dryja, J.M. Pawalek, and D.M. Albert (1978) Demonstration of tymsinase in the adult bovine uveal tract and retinal pigment epithelium. Invest. Ophthalmol. Vis. Sci., 17:511-514. Hochstein P (1983) Futile redox cycling: Implications for oxygen radical toxicity. Fundamental and Applied Toxicology, 3:215-217. Ito, S., T. Kato, K. Shinpo, and K. Fujita (1984) Oxidation of tyrosine residues in proteins by tyrosinase. Biochem. J., 222:407-411. Kable, E.P.W. and P.G. Parsons (1989)Melanin synthesis and the action of L d o p a and 3,4-dihydroxybenzylaminein human melanoma cells. Cancer Chemother. Pharmocol. 231-7. Korytowski, W., P. Hintz, R.C. Sealy, and B. Kalyanaraman (1985) Mechanism of dismutation of superoxide produced during autoxidation of melanin pigments. Biochem. Biophys. Res. Commun. 131: 659-665. Korytowski, W., B. Kalyanaraman, I.A. Menon, T. Sarna, and R.C. Sealy (1986) Reaction of superoxide anions with melanins: electron spin resonance and spin trapping studies. Biochimica Biophysica Acta, 882:145-153. Maddams, W.F. (1980)The scope and limitation of curve fitting. Applied Spectroscopy, 34:245-267. Mannagh, J . , D.V. Arya, and A.R. Irvine, Jr. (1973) Tissue culture of human retinal pigment epithelium. J. Invest. Ophthalmol., 1252-64. Marx, J . L . (1987) Oxygen free radicals linked to many diseases. Science, 235:529-531. Pawalek, J.M. and A.B. Lerner (1978) 5, 6-Dihydroxyindole is a melanin precursor showing potent cytotoxicity. Nature (London) 276:627-628. Pawalek, J.M., A. Korner, A. Bergstrom, and J. Bologna (1980) New regulators of melanin biosynthesis and the autodestruction of melanoma cells. Nature (London) 286:617-619. Prota, G. (1980) Recent advances in the chemistry of melanogenesis in mammals. J. Invest. Dermatol. 75:122-127. Rorsman, H., G. Agrup, C. Hansson, A.M. Rosengren, and E . Rosengren (1986) Detection of pheomelanins. In: Pigment Cell, S. Klaus, ed Karger, Basel, vol4, pp 244-252. Sealy, R.C., W. Puzyna, B. Kalyanaraman, and C.C. Felix (1984) Identification by electron spin resonance spectroscopy of free radicals produced during autoxidative melanogenesis. Biochemica Biophysica Acta, 800:269-276. Thompson, A., E.J. Land, M.R. Chedekel, K.V. Subbarao, and M.R. Truscott (1985) A pulse radiolysis investigation of the oxidation of the melanin precursors 3,4-dihydroxyphenylalanine(dopa) and the cysteinyldopas. Biochemica Biophysica Acta, 843:49-57.

386

K. Akeo et al.

Tomita, Y., A. Hariu, C. Kato, and M. Seiji (1984) Radical Production during tyrosinase reaction, dopa-melanin formation, and photoirradiation of dopa-melanin. J. Invest. Dermatol., 82573-576. Ueno, N. and B. Chakrabarti (1988)Monitoring in situ circular dichroism of the intact vitreous: a new approach. J. Biochem. Biophys. Methods. 1-8.

Wick, M.M., L. Byers, and E. Frei, I11 (1977) L-dopa: selective toxicity for melanoma cells in vitro. Science 197:468. Wick, M.M. (1979) Levodopa and dopamine analogs: melanin precursols as antitumor agents in experimental human and murine leukemia. Cancer Treat. Rep. 63:991-997.