Copyright © Pigment Cell Res 2000PIGMENT CELL RES 13: 80–88. 2000Printed in Ireland—all rights reser6ed ISSN 0893-5785

Original Research Article

Melanin Granules Prevent the Cytotoxic Effects of L-DOPA on RetinalPigment Epithelial Cells in Vitro by Regulation of NO and SuperoxideRadicals

KIYOSHI AKEO1, SACHI AMAKI2,3, TAIJU SUZUKI4 and TADAHISA HIRAMITSU4

1Department of Ophthalmology, Takasaki National Hospital, Gunma, Japan2Department of Ophthalmology, Kawasaki Municipal Hospital, Kanagawa, Japan3Department of Ophthalmology, School of Medicine, Keio Uni6ersity, Tokyo, Japan4Photon Medical Research Center, Hamamatsu Uni6ersity, School of Medicine, Shizuoka, Japan*Address reprint requests to Kiyoshi Akeo, M.D., Department of Ophthalmology, Takasaki National Hospital, 36 Takamatsu-cho,Takasaki-shi, Gunma, 370-0829 Japan.

Received 5 June 1999; in final form 12 January 2000

Inasmuch as the nitrogen cycle elicits the direct reduction of types of RPE cells were cultured separately in medium withN2 to NH3 through enzymatic reactions, and inasmuch as L-DOPA under an atmosphere containing 20, 10 or 5%L-DOPA (L-dihydroxyphentlalamine), a catecholamine, can oxygen. Cell numbers were counted using a Coulter counter,

and SOD activities were determined following incubation forbe a source of nitric oxide (NO), it is possible that melanin24, 48 or 72 hr using a modification of the luminol assay. Thegranules in the eye affect the generation of NO, which causes

damage to the retinal pigment epithelial (RPE) cells during results obtained indicated that: (a) NO was produced fromthe oxidation of L-DOPA. In order to confirm this possibility, L-DOPA in a concentration-dependent manner and waswe analyzed the correlations of NO generation, cell growth, trapped quantitatively by carboxy-PTIO; (b) the generation of

NO was inhibited more markedly in the melanotic cell lineand superoxide dismutase (SOD) activities in two typesthan in the amelanotic one, suggesting an increased tolerance(melanotic and amelanotic) of bovine RPE cells following

exposure to L-DOPA. NO generation from L-DOPA was to L-DOPA-derived cytotoxicity in the former; and (c) thedetermined using an NO detector that is reliant on redox SOD activities were more affected by oxygen concentration in

the melanotic cells than in the amelanotic ones. From thesecurrents. The concentration of NO was measured in terms ofresults, it is concluded that melanin granules in RPE cellsdiffusion currents run between a working electrode and ahave a role in preventing the cytotoxicity derived from L-counter electrode, both being set in culture medium placed in

a Petri dish. For the assays, L-DOPA was added to the DOPA and in regulating the generation of NO and superox-medium at various concentrations (5, 29.9, 79.4, 152.7 or 249 ide radicals.mM), and 6 min after addition, an NO-trapping agent 2,4-car-

Key words: Retinal pigment epithelium, L-DOPA, Cellboxyphenyl-4,4,5,5-tetramethylimidazole-1-oxyl 3-oxide (car-boxy-PTIO) was also added. The melanotic and amelanotic growth, Nitric oxide, SOD activities

involved in biogenesis of neurotransmitters and melanins inboth the brain and the eye. It is also known that retinalpigment epithelial (RPE) cells, when exposed to light, aresubjected to lipid peroxidation as a result of phagocytizationand digestion of photoreceptor outer segments, which arerich in polyunsaturated fatty acids (PUFA) (4–6). AlthoughRPE cells contain melanosomes, those particles are known

INTRODUCTIONIt is well known that deficiency of L-DOPA (L-dihydrox-yphentlalamine) causes degeneration of the substantia nigrain the brain and of the retina in the eye, and that L-DOPAis useful for treatment of Parkinson’s disease. Leguire et al.(1–3) reported that L-DOPA improves the visual impair-ment in an amblyopia in childhood. This situation is associ-ated with the fact that catecholamines, such as L-DOPA, are

Pigment Cell Res. 13, 200080

not to provide any direct protection against light damagefor the immediately underlying photoreceptor cells throughtheir light absorption (7). In view of the fact that bothsuperoxide anions and hydrogen peroxide are producedduring DOPA-melanin formation by tyrosinase (8, 9), itcan be said that L-DOPA in RPE cells behaves as a freeradical under highly oxidative conditions such as hyper-oxia (10) or under light exposure (11), both of whichcause cell damage. In vitro studies on bovine RPE cellsdisclosed that tyrosinase activities in these cells signifi-cantly differ between melanotic and amelanotic lines (12).We reported that in RPE cells in vitro, DOPA-oxidationaffects the procession of the cell cycle in a dose-dependentmanner, by increasing the ratio of cells arrested at the Sphase and by decreasing those at G1, and that such effectsare more pronounced in melanotic than in amelanotic cells(13).

Hollocher and Hibbs (14) reported evidence that thenitrogen cycle initiates direct reduction of N2 to NH3 bynitrogenase, which catalyses the formation of intermediatesleading to N-oxides such as NO3

−, NO2−, NO, N2O, and

N2. It is quite plausible that L-DOPA is utilized as thesource of dopamine and noradrenalin, and generates a setof the intermediates leading to N-oxides in vivo. Based oninteractions between superoxide and NO, we assume thatL-DOPA inhibits the growth of RPE cells under an atmo-sphere containing 20% oxygen, and that such inhibition islowered under 10% oxygen (10). Thus, we attempted todetermine if melanin granules have any regulatory effectson the oxidation of L-DOPA, as well as on the generationof NO and superoxide, in bovine RPE cells. This is thefirst study reporting the possible generation of NO fromL-DOPA. For determination of NO generation in vitro,we adopted a method to measure redox currents at anorder of pA using an NO meter. The total cell numbersand superoxide dismutase (SOD) activities in cultures areused as indicators of cell viability or the cytotoxicity ofreactions in these experiments.

MATERIALS AND METHODSChemicals and Instrumentation

The following reagents used for tissue culture were ob-tained from Gibco Laboratories (Grand Island, NY): fetalbovine serum (FBS); Dulbecco’s Modified Eagle’s Medium(DMEM); DMEM without phenol red (PR-DMEM);sodium bicarbonate solution (NaHCO3); penicillin andstreptomycin; trypsin; and phosphate-buffered saline freeof Ca2+ and Mg2+ (CMF-PBS). L-DOPA, bovine ery-throcyte SOD (2,990 U/mg solid), and bovine liver cata-lase (44,000 U/mg protein) were obtained from SigmaChemical Co. (St. Louis, MO). Petri dishes (diameter 35and 60 mm) and 24 multi-well (diameter 16 mm) plateswere obtained from Corning Glass Works (Corning, NY).The humidified incubator used was from VWR ScientificCo.

For measurement of the partial pressures of oxygen,pO2, a diaphragm electrode (polarograph) was used, whichconsisted of a platinum (Pt) cathode and a chlorinatedsilver anode, following the installation in an oxygen-regu-

lated incubator (B 5060 EK/O2; Heraeus, Hanau, Ger-many). The pO2 in the incubator was calibrated by air,i.e., 21.0% oxygen, which was blown into the electrode.Cell counting was carried out using a Coulter counter(Coulter Electronics Ltd., Luton, Beds, UK). For determi-nation of NO, 2-4-Carboxyphenyl-4,4,5,5-tetramethylimi-dazole-1-oxyl 3-oxide (Carboxy-PTIO) (Dojindo Lab-oratories Co., Kumamoto, Japan) and an NO meter(NO-501, Inter Medical Co., Tokyo, Japan) were used.The NO meter used was designed to measure pA-orderredox currents: the working electrode was made of a 0.2mm Pt/Ir (iridium) alloy (Pt 90%, Ir 10%) wire coatedwith three layers of membranes, each composed of KCl,NO-selective resin, and pure silicone (15), whereas thecounter electrode was made of carbon fiber. The formerwas made as follows: first, a KCl membrane was electro-chemically coated onto a Pt/Ir wire; second, NO-selectiveresin was coated onto the wire by immersion of the wirethus prepared in 0.8% nitrocellulose solution, followed bydrying under air for 12 hr; and finally, the wire was thencoated with silicone by rinsing it in nitrocellulose solutionprepared by dissolving commercially available collodion(Wako Pure Chemical Co., Tokyo, Japan) 5 times in di-ethylether:ethanol (3:1). A KCl membrane was installedfor suppression of an over-voltage raised by NO discharge,and a silicone membrane was installed in the outermostlayer to avoid a non-specific ionic effect and electrochemi-cal reactions. The working electrode thus prepared yielded+0.4 to +0.8 V as a result of the electrochemical oxida-tion of NO. Such polarographic currents were detectedwith a current voltage converter equipped in a high inputimpedance preamplifier set near the electrode pair.

For the SOD assay, Tris–HCl, diethylenetriaminepen-taacetic acid (DETAPAC; an iron chelator), bovine serumalbumin (BSA), hypoxanthine, luminol, xanthine oxidase(type I 50U/1.2 ml) (Sigma Chemical Co., St. Louis, MO)and a luminometer (Luminescence Reader, Aloka Co.,Tokyo, Japan) were used. An assay of proteins was madeaccording to Smit et al. (17) using bicinchonic acid (BCA)and CuSO4.

Culture of Bovine Melanotic and Amelanotic RPE Cells

Following aseptic dissection of the anterior segments ofthe bovine eyes, the retina and the vitreous were removedfrom each eye cup. Melanotic and amelanotic bovine RPEcells were collected separately in 60 mm Petri dishes byscraping Bruch’s membrane with a round-tipped Pasteurpipette under a dissecting microscope. These RPE cellswere cultured in 95% air/5% CO2 in a humidified incuba-tor at 37°C and maintained in a DMEM supplementedwith 15% FBS, 0.125% NaHCO3, penicillin 41.5 units/ml, and streptomycin 41.5 mg/ml (hence completeDMEM). Culture media were changed with freshly-pre-pared complete DMEM weekly until the cells reachedconfluence. RPE cells were dislodged with 2.5% trypsin,centrifuged in centrifuge tubes and replated for subcultureusing 2-week-old primary cultures. Porcine RPE cells werealso prepared by a similar procedure for supplemental ex-periments.

Pigment Cell Res. 13, 2000 81

Detection of NO in the Culture Media

A set of working and counter electrodes supported in shortmicropipette tips was immersed in a Petri dish (diameter35 mm). After replacing a complete PR-DMEM with PR-DMEM and subsequent complete shielding of the dishwith Parafilm, a thin Teflon tube placed above the dishwas connected with a tube from a micropipette in the dishcontaining RPE cells. This system was set up to mix L-DOPA in PR-DMEM quickly without vibrations by stir-ring, because minor mechanical changes on the electrodessuch as moving their position or changing the flow rateyielded spurious currents as a result of the high sensitivityof the system. After achieving a stable basic current, L-DOPA was added to the medium so as to adjust its con-centration to 5, 29.9, 79.4, 152.7 or 249 mM every 2 min.Six minutes after the addition of 249 mM L-DOPA, Car-boxy-PTIO (249 mM) was added to the media. Currentchanges were monitored throughout the experiment.

Counting of Cultured RPE Cells in Vitro

RPE cells at the third passage were prepared by subcultur-ing the primary culture in a stepwise fashion over 4 weeks.The number of third passage cells was adjusted to about5,000–10,000 cells/ml by addition of complete DMEM.One milliliter aliquots of the cell suspension were placed ina triplicate of 24 multi-well plates for each experimentcarried out for a given condition. Twenty-four hours afterplating, three wells were provided for determination of thezero-time cell count under a Coulter counter. The concen-trations of L-DOPA were adjusted to either 100 or 250 mMin PR-DMEM containing the same concentrations of FBSand antibiotics as in the complete DMEM. After the oxy-gen concentration in the incubators was adjusted to 20, 10or 5% by the addition of nitrogen, RPE cells were culturedin the PR-DMEM with L-DOPA under a given condition.Because each experiment was carried out using two incu-bators, it was necessary to periodically confirm that thedifferences between the incubators came from oxygen con-centrations and not from some other unrecognized factors.Therefore, duplicate cultures were made in 20% oxygenand hypoxia (10 or 5% oxygen) and total cell numberswere determined using a Coulter counter after 24, 48, and72 hr. At the same time, we carried out experiments inwhich SOD (90 U/ml) and catalase (900 U/ml) were addedto the medium, because we previously found that in cul-tured RPE, the addition of these agents reduced the celldamage that occurred following exposure to either 100 or250 mM L-DOPA (10).

SOD Assays on RPE Cells in Vitro

For this purpose, we used bovine melanotic and amelan-otic RPE cells at the third passage. Porcine melanotic RPEcells at the same passage were also examined for compari-son (data not shown). After 48 or 72 hr exposure todifferent oxygen levels (20, 10 or 5% oxygen) and to differ-ent concentrations of L-DOPA (100 or 250 mM), bovineand porcine RPE cells were trypsinized, centrifuged intopellets and stored at −20°C. For the SOD assay, 20 mM

Tris–HCl buffer with 0.1 mM DETAPAC was added tothe pellets, and the cells were lysed by application of sev-eral cycles of freezing and thawing. The activities of SODwere measured using a modification of the luminol assaydescribed by Bensinger and Johnson (16), which countedthe emission of photons of light emitted from luminol byits interaction with the superoxide anion. The reaction pro-ceeded in the assay buffer containing 100 mg/ml BSA, 70mg/ml hypoxanthine and 10 mM luminol. One hundredmicroliters of a SOD standard solution or cell lysate wasadded to mini-scintillation vials containing 1 ml of reactionmixture. To start the assay, 100 ml of xanthine oxidase(0.54 U/ml) was added, rapidly mixed, and placed in thesample chamber of a luminometer. Photon emission wasmeasured repeatedly at 6-s intervals to obtain the maxi-mum emission value, which was expressed as a percentageof the control emitted by a mixture without SOD. Thenanogram equivalents of SOD were determined from astandard curve showing suppression of photon emission bythe amount of bovine erythrocyte SOD added. Proteinconcentrations in aliquots of each sample were determinedusing the BCA method. The specific activity of SOD wasexpressed as nanogram equivalents of SOD per mgprotein.

Statistical Analysis

Data from duplicate experiments were combined and ex-amined by pairwise multiple comparison procedures(Tukey test) (18) to identify which experimental conditionsproduced significant changes. A level of P50.05 wastaken to be statistically significant. The cell numbers andSOD activities obtained under three different oxygen con-centrations were combined, and the standard errors werecalculated from three different experiments as given by thebars in the figure columns. Because we were using anoxygen-regulated incubator, it was possible to carry outthe experiments under two different concentrations of oxy-gen at one time. Therefore, the data obtained using either5 or 10% oxygen were compared with those using 20%oxygen that was a standard for tissue cultures.

To compare the responses of RPE cells with and with-out melanin granules to L-DOPA and oxygen, we adoptedan analysis of the stepwise linear regression (19) for thedata obtained from two cell types of melanotic and ame-lanotic RPE cells. We used a model that includes differentterms for cell type-specific differences regarding the diffu-sion currents, cell numbers and SOD activities:

Diffusion current i=ba1+ba1(DOPA)i+ba2(IPE)i+ba3(IPE)(DOPA)i+Ei. (1)

Log10 (cell number)i=bb0+bb1(time)i+bb2(DOPA)i+bb3(oxygen)i+bb4(DOPA)(oxygen)i+bb5(IPE)i+bb6(IPE)(DOPA)i+bb7(IPE)(oxygen)i

+bb8(IPE)(DOPA)(oxygen)+Ei. (2)

SOD activities i=bc0+bc1(DOPA)i+bc2(oxygen)i+bc3(DOPA)(oxygen)i+bc4(IPE)i+bc5(IPE)(DOPA)i+bc6(IPE)(oxygen)i+bc7(IPE)(DOPA)(oxygen)+Ei. (3)

Pigment Cell Res. 13, 200082

IPE, an index for melanin granules, was set to 0 and 1 foramelanotic and melanotic cells, respectively. This indicatorwas used for another variable to calculate PE-specific dif-ferences in relation to a specific factor, as seen in(IPE)(DOPA) taken as a response of melanotic cells toDOPA. Diffusion current, log10 (cell number), and SODactivities in an individual plate (i) used for each experi-ment was considered to be a possible function of thenumber of plates (ba0), the incubation time of DOPA, theconcentrations of DOPA and oxygen, the interaction be-tween DOPA and oxygen, and the random error term(Ei). These experimental data were also analyzed usingcomputer programs that generated the values, the signifi-cance of ba0, ba1, ba2, ba3, ba4, ba5, ba6, ba7, and ba8,and that determined how well the model fits the data tothe concentration of DOPA (0, 100, and 250 mM) and ofoxygen (5, 10, and 20%) (r2; a level of the significance incomposite model). We tested the hypothesis that b=0against the alternative b"0, where b is the regressioncoefficient of the population, using the following assump-tions: A1 [for each fixed variable x, i.e., (DOPA), (oxy-gen), (DOPA)(oxygen), (IPE)(DOPA), (IPE)(oxygen),(IPE)(DOPA)(oxygen), the random variable g, i.e.,Log10(cell number), SOD, and diffusion current, is normalwith mean and variance s2 where the latter is independentof x] and A2 [i, performances of the experiments by whichwe obtained a sample (x1, g1), … , (x1, g1) were indepen-dent]. We determined the significant level P from a tableof F-distributions with degrees of freedom whether thehypothesis was rejected or not.

Multiple linear regression analysis (19) was commonlyused to forecast values for a given variable (dependentvariable) based upon the values of other variables (inde-pendent variables). We estimated the relative contributionsof DOPA (100 and 250 mM), and catalase (900 U/ml) andSOD (90 U/ml) (CS), as independent variables, to therepression of Log10 (cell number)i, as dependent variables.The model for analysis was:

Log10(cell number)i=bd0+bd1(time)i+bd2(DOPA)i+bd3(CS)i+bd4(DOPA)(CS)i+Ei. (4)

The statistical package used in this analysis was SigmaStat (Jandel Scientific Co., San Rafaero, LA).

RESULTSGeneration of NO from L-DOPA in a Cell-free CultureMedium

Addition of L-DOPA at concentrations of 5, 29.9, 79.4,152.7 or 249 mM to the cell-free culture medium increasedthe diffusion current in a dose-dependent fashion (Fig.1A). This was considered to indicate NO generation, be-cause the addition of Carboxy-PTIO (249 mM) to theabove medium decreased diffusion currents within 2.5 minby up to 64% of the current observed in the presence of249 mM L-DOPA (Fig. 1B).

Generation of NO from L-DOPA in Medium Incubatedwith Bovine Melanotic and Amelanotic RPE Cells

The changes in diffusion currents resulting from the addi-tion of L-DOPA to the culture medium containing eitherbovine melanotic or amelanotic RPE cells in vitro aregiven in Fig. 2. As shown, addition of 249 mM L-DOPAto the medium yielded diffusion currents of 1.153 nA formelanotic cells and of 1.347 nA for amelanotic cellswithin 2 min of addition, both being lower than those(1.67 nA) observed under cell-free conditions. The diffu-sion currents derived from added L-DOPA were appar-ently lower in a medium with melanotic RPE cells thanwith amelanotic ones, indicating a detectable degree ofinhibition of the generation of NO in the former (Table1).

Fig. 1. Changes in diffusion currents generated by addition ofL-DOPA (A) or Carboxy-PTIO (249 mM) (B) to cell-free culturemedia under air (20% oxygen).

Pigment Cell Res. 13, 2000 83

Fig. 2. Changes in diffusion currents generated by addition ofdifferent concentrations of L-DOPA to the culture media containingmelanotic or amelanotic RPE cells under air (20% oxygen). Marksfor bars: open, cell-free; stippled, with amelanotic RPE cells;shaded, with melanotic RPE cells. Abbreviations: N.S.=no signifi-cant difference; * PB0.05; ** PB0.01; *** PB0.001.

Fig. 3. Effects of L-DOPA (250 mM) and oxygen on the growth ofmelanotic (A) and amelanotic (B) RPE cells after 72 hr incubation.Note a significant difference in cell numbers between two types ofcells incubated under 5% oxygen. Marks for oxygen concentrations:open, 5%; stippled, 10%; shaded, 20%. For abbreviations of thestatistics refer to the legend in Fig. 2.

Effects of L-DOPA and Oxygen on the Growth of RPECells

When melanotic and amelanotic types of bovine RPE cellswere cultured separately over 48 hr in medium containingeither 100 or 250 mM L-DOPA under different oxygenconcentrations (5, 10 or 20%) (Fig. 3), a trend was ob-served that at a given concentration of L-DOPA, loweringthe oxygen concentration prevented a decrease of cellgrowth (cell numbers per culture), and that such effectsbecame more distinct at a higher concentration (250 mM)of L-DOPA. In other words, the growth (survival) of RPEcells was adversely affected in a dose-dependent fashionby the presence of L-DOPA with an increase of oxygenpressure. These results were observed more distinctly inthe amelanotic type cells than in the melanotic ones, sug-gesting the role of melanin granules for cell growth (sur-vival) in the presence of L-DOPA and oxygen. Fig. 3indicates the effects of oxygen concentrations on the cellgrowth of melanotic and amelanotic cells after 72 hr incu-bation. The statistical analysis (Table 2) showed that suchinteraction apparently inhibited the growth of both melan-otic and amelanotic RPE cells [coefficient value of(DOPA)(oxygen)= −0.000451, PB0.001], and that sig-nificant differences exist between melanotic and amelanoticRPE cells with respect to cytotoxic effects caused by aninteraction of L-DOPA and oxygen. The latter would im-ply that melanotic RPE cells are significantly more toler-

ant to cytotoxicity caused by DOPA–oxygen interactionthan amelanotic RPE cells [coefficient value of(IPE)(DOPA)(oxygen)=0.000125, PB0.001]. This also in-dicates that in hypoxia, an increase in L-DOPA concentra-tion brings about an increase in RPE cell growthregardless of the presence or absence of melanin granules[coefficient value of (DOPA)=0.00316, PB0.001], andthat such cell growth was significantly higher in amelan-otic RPE cells than in melanotic ones [coefficient value of(IPE)(DOPA)= −0.00103, PB0.02].

Table 1. Effects of NO generation from L-DOPA on the growth ofmelanotic RPE cells

Diffusion currents=ba0+ba1(DOPA)+ba2(IPE)i+ba3(DOPA)(IPE)

(DOPA)(IPE)(DOPA)ba3ba1 Fit of the model (n=24)

r2=0.817Value −2.0366.7130.0010.0020.001P

Pigment Cell Res. 13, 200084

Effects of SOD and Catalase on Cell Growth in thePresence of 20% Oxygen

The growth of amelanotic RPE cells, when cultured in thepresence of 250 mM L-DOPA for 24 hr (Fig. 4A), 48 hr (Fig.4B), and 72 hr (Fig. 4C) without the addition of SOD andcatalase, was lower than that of melanotic RPE cells. Whenthese two anti-oxidative enzymes were added to the abovemedium, the cell growth expressed in terms of cell numbersper culture were larger in melanotic RPE cells than inamelanotic ones at 48 hr (Fig. 4B) and 72 hr (Fig. 4C). Ananalysis of multiple linear regression showed that the growthof amelanotic RPE cells [coefficient value of (DOPA)= −0.00634, PB0.001; coefficient value of (DOPA)(CS)=0.00594, PB0.001] was more affected by exposure toL-DOPA in 20% oxygen than melanotic RPE cells [coeffi-cient value of (DOPA)= −0.00395, PB0.001], and that areduced cell growth was recovered by the addition of anti-oxidative enzymes more distinctly in amelanotic cells than inmelanotic ones [coefficient value of (DOPA)(CS)=0.00374,PB0.001] (Table 3).

Effects of L-DOPA and Oxygen on SOD Activities

The SOD activities in melanotic and amelanotic types ofRPE cells were increased by the addition of 250 mM L-DOPA [coefficient value of (DOPA)=3.529, PB0.003],except in the case of melanotic RPE cells under 5% oxygen(Table 4Fig. 5). The SOD activities of amelanotic RPE cellswere less under 5% oxygen than under 10 or 20% oxygen.However, the SOD activities of melanotic RPE cells inmedium without L-DOPA under 5% oxygen were signifi-cantly higher than under 20% oxygen (Fig. 5). The decreas-ing rate in the SOD activities responding to increasedoxygen concentrations was larger in melanotic RPE cellsthan in amelanotic ones. Statistical analysis disclosed thatthere were significant differences with respect to the re-sponses of melanotic and amelanotic RPE cells on exposureto L-DOPA and oxygen disregarding variations in the SODactivities [coefficient value of (IPE)(oxygen)= −60.041, PB0.05] (Table 4).

DISCUSSIONWe successfully detected the generation of NO from L-DOPA by measuring the diffusion currents in combinationwith an NO monitor and carboxy-PTIO, a selective scav-enger of NO that inhibited them up to 64% at the sameconcentration of L-DOPA (20). Because the amplitude ofdiffusion currents was different in each experiment, depend-ing on the sensitivity of the electrode, we used one electrodefor one set of experiments carried out at the same tempera-

ture, pH, and oxygen concentration. Ichimori et al. (15)reported that the electrode current produced by 50 mM ofS-nitroso-N-acetyl-DL-penicillamine (SNAP) in an in vitroNO-measuring system corresponds to a 20 pA current at 70nM NO, and that it was suppressed by an NO scavengersuch as carboxy-PTIO (30 mM). They stated that carboxy-PTIO produced 20% of the current induced by 0.1 mMSNAP. Such a current may come from attachment of com-pounds yielded by L-DOPA and carboxy-PTIO to the elec-trode surface (15) or from NO that is not scavenged bycarboxy-PTIO at this concentration. We speculate that 36%of the current produced by L-DOPA is attributable to NO,and that 64% of the current is a non-specific unidentifiedcurrent.

We reported that L-DOPA stimulated the growth of RPEcells in vitro under 10% oxygen (10). Addition of L-DOPAduring low oxidative stress resulted in an increase of RPEcells at the G2+M phase (13) and an increase in theribonucleic acid to deoxyribonucleic acid (RNA:DNA) ratio(21). In those studies, we demonstrated that L-DOPA playsa role in the maintenance of the viability of RPE cellsthrough NO derived from the oxidation of amines. Fromthese findings, it was expected that L-DOPA might be usefulin the treatment of retinal degeneration by its administrationto patients who were suffering from retinal degenerationcaused either by low oxidative stress conditions or in theabsence of exposure to light and oxygen. On the other hand,it is known that L-DOPA causes cytotoxic effects on RPEcells in the retina under conditions of high oxidative stresssuch as exposure to light or hyperoxia. Presumably at pH7.4, the likely pH of the culture medium, L-DOPA under-goes autoxidation yielding superoxide and hydrogen perox-ide, both of which are capable of reaction withnitrogen-containing compounds possibly forming NO. Wecould speculate that exposure of the retina to light orhyperoxia probably causes increased formation of superox-ides that subsequently scavenge NO from L-DOPA andproduce peroxynitrite. Because the nitrogen cycle initiatesthe direct reduction of N2 to NH3 by nitrogenase (14), theenzymatic reactions would yield a set of NO intermediatessuch as NO3

−, NO2−, NO, N2O, and N2. We wish to

emphasize that this study is the first to indicate that L-DOPA has two different effects on RPE cells: excitation andinhibition.

When L-DOPA is converted to quinolic derivatives in20% oxygen, superoxide is produced during the process ofmelanin formation. NO combined with superoxide formsperoxynitrite, which leads to the formation of hydroxylradicals (22). We speculate that these hydroxyl radicalscould induce lipid peroxidation in cell membranes, thus

Table 2. Effects of L-DOPA and oxygen on the growth of melanotic and amelanotic RPE cells

Log10 (number of cells)=bb0+bb1(time)i+bb2(DOPA)i+bb3(oxygen)i+bb4(DOPA)(oxygen)i+bb5(IPE)i+bb6(IPE)(DOPA)i+bb7(IPE)(oxygen)i+bb8(IPE)(DOPA)(oxygen)

(DOPA) (DOPA)(oxygen) (IPE)(DOPA) (IPE)(DOPA)(oxygen)(Time)bb6bb4bb2 bb8bb1 Fit of the model (n=198)

0.00807 0.00316 −0.000451 −0.00103 0.000125 r2=0.845Value0.001 0.020.0010.001 0.001P 0.001

Pigment Cell Res. 13, 2000 85

Fig. 4. Effects of SOD and catalase on the growth of melanotic and amelanotic RPE cells after incubation for (A) 24, (B) 48 or (C) 72 hrin the presence of different concentrations of L-DOPA under 20% oxygen. Symbols for cell types: open, amelanotic RPE cells; solid,melanotic RPE cells. Symbols for experimental conditions: triangles, with SOD and catalase; circles, without SOD and catalase. Note thedifference in cell numbers between the experiments carried out with and without SOD and catalase. For abbreviations of the statistics referto the legend in Fig. 2.

damaging RPE cells. This hypothesis would support thenotion that cell damage by L-DOPA results from peroxyni-trite rather than from superoxide alone. The fact that NO isinactivated by superoxide in hyperoxia is another importantreason for the injury caused to RPE cells. The concentra-tions of the combined addition of SOD and catalase used(90 and 900 units/ml, respectively) are selected on the basisof previous experiments that showed that these concentra-tions are sufficient to ensure the improvement of the reducedgrowth and/or survival of RPE cells by L-DOPA (10, 11).SOD scavenges superoxides and hydroperoxides, and alsometabolizes peroxynitrites by phenolic nitration (23). Inthese studies, addition of L-DOPA to the medium increasedthe SOD activities of RPE cells, regardless of the presenceor absence of melanin granules, and we realized that theconcentrations of L-DOPA added to the medium correlatedwith the quantity of superoxide-induced peroxynitrite.

When the experiments were carried out using cell-freePetri dishes sealed with Parafilm, NO generation was foundto be dependent on the concentration of L-DOPA. In thepresence of cells, the diffusion current was diminished (Fig.

2) by up to 30%. The presence of melanin granules in themelanotic RPE cells inhibited the generation of NO byaddition of L-DOPA because these RPE cells were able totake up more L-DOPA than amelanotic ones. Whilst someof the influence on the diffusion current appeared to beunrelated to pigmentation, there was evidence indicating alink with melanin or melanogenesis. RPE cultures exposedto three differing oxygen levels for 24, 48 or 72 hr (Fig. 3)

Table 3. Effects of SOD and catalase on the growth of melanoticand amelanotic RPE cells incubated with L-DOPA

Log10 (number of cells)=bd0+bd1(DOPA)+bd2(DOPA)(CS)

(DOPA)(CS)(DOPA) Fit of the modelbd2 (n=54)bd1

Melanotic RPE cellsValue −0.00395 0.00374 r2=0.521

0.0010.001 0.001P

Amelanotic RPE cellsValue r2=0.6910.00594−0.00634

0.001P 0.001 0.001

Pigment Cell Res. 13, 200086

Table 4. Effects of L-DOPA and oxygen on SOD activities inmelanotic and amelanotic RPE cells

SOD activities=bc0+bc1(DOPA)+bc6(IPE)(oxygen)

(DOPA) (IPE)(DOPA)bc1 bc6 Fit of the model (n=48)

Value r2=0.2613.529 −60.041P 0.003 0.0050.05

CONCLUSIONSWe tried to detect NO generated by L-DOPA using theprinciple of the NO meter based on measurement of redoxcurrents between the working and the counter electrodes.NO was produced in a concentration-dependent fashionfrom L-DOPA and was scavenged by carboxy-PTIO.Melanotic RPE cells inhibited the generation of NO byL-DOPA, and were more tolerant to cytotoxicity byDOPA–oxidation interaction than amelanotic ones. TheSOD activity of melanotic RPE cells was more affected byoxygen than that of amelanotic ones. Melanin granulesprevented the cytotoxicity of DOPA–oxidation interactionbecause the generation of NO and superoxide radicals, i.e.,peroxynitrite, could be regulated by them.

Acknowledgements – A part of this study was presented at the103rd Japanese Ophthalmic Society Meetings in Chiba City onApril 21 and Association for Research in Vision and Ophthalmol-ogy on May 10 in Fort Lauderdale, in 1999.

REFERENCES1. Leguire LE, Walson PD, Rogers GL, Bremer DL, McGregor

ML. Levodopa/carbidopa treatment for amblyopia in olderchildren. J Pediatr Ophthalmol Strabismus 1995;32:143–151

2. Leguire LE, Walson PD, Rogers GL, Bremer DL, McGregorML. Longitudinal study of levodopa/carbidopa for childhoodamblyopia. J Pediatr Ophthalmol Strabismus 1993;30:354–360

3. Leguire LE, Walson PD, Rogers GL, Bremer DL, McGregorML. Levodopa/carbidopa for childhood amblyopia. InvestOphthalmol Vis Sci 1993;34:3090–3095

4. Poincelot RP, Abrahamson EW. Fatty acid composition ofbovine rod outer segments and rhodopsin. Biochim BiophysActa 1970;202:382–385

5. Hendricks TK, Klompmakers AA, Daemen FJM, Bonting SL.Biochemical aspects of the visual process. XXII. Movement ofsodium ions through bilayers composed of retinal and rod outersegment lipids. Biochim Biophys Acta 1976;443:271–281

6. Stone WL, Farnsworth CC, Dratz EA. A reinvestigation offatty acid content of bovine, rat and frog retinal rod outersegments. Exp Eye Res 1979;28:387–397

7. LaVail MM, Gorrin GM. Protection from light damage byocular pigmentation: Analysis using experimental chimeras andtranslocation mice. Exp Eye Res 1987;44:877–889

8. Searly RC, Puzyna W, Kalyanraman B, Felix CC. Identifica-tion by electron spin resonance spectroscopy of free radicalsproduced during autoxidative melanogenesis. Biochim BiophysActa 1984;800:269–276

9. Tomita Y, Hariu A, Kato C, Seiji M. Radical productionduring tyrosinase reaction, dopa-melanin formation, and pho-toirradiation of dopa-melanin. J Invest Dermatol 1984;82:573–576

10. Akeo K, Ebenstein DB, Dorey CK. Dopa and oxygen inhibitproliferation of retinal pigment epithelial cells, fibroblasts andendothelial cells in vitro. Exp Eye Res 1989;49:335–346

11. Dorey CK, Delori FC, Akeo K. Growth of cultured RPE andendothelial cells is inhibited by blue light but not green or redlight. Curr Eye Res 1990;9:549–559

12. Basu PK, Sarkar P, Menon I, Carre F, Persad S. Bovine retinalpigment epithelial cells cultured in vitro: Growth characteris-tics, morphology, chromosomes, phagocytosis activity, tyrosi-nase activity, and effect of freezing. Exp Eye Res1983;36:671–683

13. Akeo K, Tanaka Y, Okisaka S. A comparison between melan-otic and amelanotic retinal pigment epithelial cells in vitroconcerning the effects of L-dopa and oxygen on cell cycle.Pigment Cell Res 1994;7:145–151

14. Hollocher TC, Hibbs JB Jr. The nitrogen cycle. In: Feelisch M,Stamler JS. Methods in Nitric Oxide Research. Part III.Chichester: John Wiley & Sons; 1996. pp. 120–121

exhibited DOPA-dependent reduction in cell numbers thatwas proportional to the oxygen concentration. At thehighest oxygen exposure (20%), there was an apparent re-duction in cell numbers in both melanotic and amelanoticRPE cultures, although this was more pronounced in ame-lanotic cells. The deduction from this finding was consistentwith oxidative damage to cells as a result of DOPA autoxi-dation products, and that this was partially abrogated bysome properties of pigmented cells, possibly radical scaveng-ing by melanin. Another possibility was that melanotic RPEcells have less SOD activities or produce more superoxidesthan amelanotic ones, i.e., there would be more superoxidesto inactivate NO in the medium of melanotic RPE cells. Inthese studies, we demonstrated that the SOD activity ofmelanotic RPE cells incubated with L-DOPA is more af-fected by oxygen than that of amelanotic ones. Peroxynitritewas generated from L-DOPA as well as from the compound3-morpholinosydnonimine (SIN-1), which releases superox-ide and NO simultaneously. Exposure of PC12 pheochro-mocytoma cells to peroxynitrite diminished their ability tosynthesize DOPA without apparent cell death. The inhibi-tion in DOPA synthesis by SIN-1 was abolished whensuperoxide was scavenged by the addition of SOD (24). Wefirst detected the generation of NO from L-DOPA itself, andwe suspect that the DOPA synthesized in PC12 pheochro-mocytoma cells might actually be the source of superoxideand NO.

Fig. 5. Effects of L-DOPA and Oxygen on SOD Activities. Forabbreviations of the statistics refer to the legend in Fig. 2. Marksfor oxygen concentrations: open, 5%; stippled, 10%; shaded, 20%.

Pigment Cell Res. 13, 2000 87

15. Ichimori K, Ishida H, Fukahori M, Nakazawa H, MurakamiE. Practical nitric oxide measurement employing a nitric oxide-selective electrode. Rev Sci Instrum 1994;65:2714–2718

16. Bensinger RE, Johnson CM. Luminol assay for superoxidedismutase. Anal Biochem 1981;116:142–145

17. Smith PK, Krohn RI, Hermanson GT, Mallia AK, GartnerFH, Provenzano MD, Fujimoto EK, Goeke NM, Olson BJ,Klenk DC. Measurement of protein using bicinchoninic acid.Anal Biochem 1985;150:76–85

18. Snedecor GW, Cochran WG. Two-way classification. In: Statis-tical Methods, 8th edition. Iowa: Iowa State University Press;1989, pp. 254–272.

19. Snedecor GW, Cochran WG. Multiple linear regression. In:Statistical Methods, 8th edition. Iowa: Iowa State UniversityPress; 1989, pp. 333–373.

20. Akaike T, Yoshida M, Miyamoto Y, Sato K, Kohno M,Sasamoto K, Miyazaki K, Ueda S, Maeda H. Antagonisticaction of imidazolineoxyl N-oxides against endothelium-derived

relaxing factor/NO through a radical reaction. Biochemistry1993;32:827

21. Kanda T, Akeo K, Murakami A, Karasawa Y, Okisaka S. Theeffects of dopa and oxygen on RNA concentrations in culturedembryonal retinal pigment epithelial cells. J Jpn OphthalmolSoc 1995;99:1123–1126

22. Beckman JS, Beckman TW, Chen J, Marshall PA, FreemanBA. Apparent hydroxyl radical production by peroxynitrite:Implications for endothelial injury from nitrite oxide and super-oxide. Proc Natl Acad Sci USA 1990;87:1620–1624

23. Beckman JS, Ischiropoulos H, Zhu L, van der Woerd M, SmithC, Chen J, Harrison J, Martin JC, Tsai M. Kinetics of superox-ide dismutase- and iron catalyzed nitration of phenols byperoxynitrite. Arch Biochem Biophys 1992;298:438–445

24. Ischiropoulos H, Duran D, Horwitz J. Peroxynitrite-mediatedinhibition of DOPA synthesis in PC12 cells. J Neurochem1995;65:2366–2372

Pigment Cell Res. 13, 200088