Free Radical Biology & Medicine, Vol. 36, No. 9, pp. 1144 –1154, 2004Copyright D 2004 Elsevier Inc.

Printed in the USA. All rights reserved0891-5849/$-see front matter

doi:10.1016/j.freeradbiomed.2004.01.019

Original Contribution

EFFECTS OF MELANIN AND MANGANESE ON DNA DAMAGE AND

REPAIR IN PC12-DERIVED NEURONS

VASYL SAVA,*,y DIANA MOSQUERA,*,y SHIJIE SONG,*,y

FERNANDO CARDOZO-PELAEZ,z and JUAN R. SANCHEZ-RAMOS*,y

*Department of Neurology, University of South Florida, Tampa, FL, USA; yResearch Services, James A. Haley Veterans Hospital,Tampa, FL, USA; and zDepartment of Pharmaceutical Sciences, Center for Environmental Health Sciences, Missoula, MO, USA

(Received 17 October 2003; Revised 7 January 2004; Accepted 23 January 2004)

Ad

E. Ellis

Univer

FL 336

Abstract—The mechanism of neurotoxicity produced by the interaction of melanin with manganese was investigated in

PC12-derived neuronal cell cultures. The cells were incubated with melanin (25–500 Ag/ml), MnCl2 (10 ng/ml–100 Ag/ml), and a combination of both substances for 24 and 72 h. Incubation with either toxicant alone resulted in a minimal

decrease in cell viability. The combination of melanin and manganese caused significant (up to 60%) decreases in

viability of PC12 cells in a dose-dependent manner. Increases in oxidative DNA damage, indicated by levels of 8-

hydroxy-2Vdeoxyguanosine (8-oxodG), was associated with decreased cell viability. Melanin alone, but not manganese

alone, resulted in increased oxidative DNA damage. The maximal increase in 8-oxodG caused by melanin was about

seven times higher than control after 24 h of exposure. The activity of the DNA repair enzyme, 8-oxoguanine DNA

glycosylase (OGG1), was increased in cells incubated with single toxicants and their combinations for 24 h. On the third

day of incubation with the toxicants, activity of OGG1 declined below control levels and cell viability significantly

decreased. Melanin was observed to have an inhibitory effect on OGG1 activity. Study of the regulation of OGG1

activity in response to melanin and manganese may provide insights into the vulnerability of nigral neurons to oxidative

stress in Parkinson’s disease. D 2004 Elsevier Inc. All rights reserved.

Keywords—DNA damage and repair, 8-Oxoguanine DNA glycosylase, Neuronal degeneration, Free radicals

INTRODUCTION

Occupational heavy metal exposure, particularly to man-

ganese, is a risk factor for parkinsonism, a syndrome of

muscle rigidity, slowness of movement, tremor, and

postural instability. The manifestations of chronic man-

ganese poisoning were noted to resemble the signs and

symptoms of idiopathic Parkinson’s disease (IPD) [1–4].

The pathological process in IPD involves the selective

degeneration of nigrostriatal dopaminergic neurons,

which results in depletion of striatal dopamine (DA)

[5]. Manganese is a potential DA-ergic neurotoxin in

dress correspondence to: Dr. Juan R. Sanchez-Ramos, The Helen

Professor of Neurology, Department of Neurology (MDC 55),

sity of South Florida, 12901 Bruce B. Downs Boulevard, Tampa,

12. Fax: +1-813-974-7200. E-mail: jsramos@hsc.usf.edu.

1144

vivo and in vitro. DA levels were reported to be sig-

nificantly decreased in the corpus striatum of monkeys

intoxicated with manganese [6]. Injection of manganese

into the rodent striatum reduced the concentration of DA

and impaired oxidative metabolism [7–9]. The mecha-

nism by which this metal injures neurons of the basal

ganglia remains unclear, but several hypotheses have

been put forth. Manganese may increase oxidative stress

by catalyzing DA autoxidation and, in the process,

generating both oxygen free radicals and toxic quinones

[10]. Another concept is that manganese triggers cell

death by an apoptotic mechanism (programmed cell

death) similar to that produced by the DA-ergic neuro-

toxin MPTP, and that is also hypothesized to occur in

IPD [11–14]. The oxidative stress and apoptosis hypoth-

eses are linked by the shared mechanism of mitochon-

drial dysfunction, which results in both oxidative stress

Melanin/manganese effects on DNA damage/repair 1145

and apoptosis. Both Mn2+ and MPP+, the toxic metabo-

lite of MPTP, cause dysfunction of mitochondrial elec-

tron transport by inhibiting Complex II and Complex I,

respectively [15–17]. Unlike the neurotoxic effect of

MPTP, which is highly selective for DA neurons, the

toxicity of manganese extends to the corpus striatum and

pallidum, resulting in a more complex clinical phenotype

that often includes dystonia and psychiatric disturbances

[9,18,19].

Catecholaminergic neurons, such as DA-ergic neurons

of the substantia nigra (SN) and noradrenergic neurons of

the locus coeruleus, accumulate neuromelanin within the

cytoplasm [20]. Neuromelanin is produced by catechol-

amine autoxidation and is unlike skin melanin (eumela-

nin), which is produced by enzymatic action of

tyrosinase [21]. The autoxidative process of neuromela-

nin formation generates reactive oxygen species that

cause cellular dysfunction by their reactions with lipids,

proteins, and nucleic acids. In addition, autoxidation of

DA forms toxic semiquinone and quinone species, which

injure cells through inhibition of sulhydryl enzymes and

reaction with nucleophilic groups within the cells [10]. In

IPD, the neuromelanin content of surviving DA neurons

in SN is significantly diminished compared with age-

matched control cases [20]. This may be because the

most heavily melanized DA neurons degenerate in IPD.

On the other hand, there is evidence that the most highly

melanized DA-ergic neurons are the last to degenerate

[22]. Neuromelanin has a high affinity for heavy metals

and accumulates endogenous as well as exogenous

metals [21]. Such a relationship can be either neuro-

protective or deleterious, depending on the extent to

which the chelated metals are prevented from catalyzing

the formation of reactive oxygen species.

It has been hypothesized and demonstrated that DNA

damage and repair processes have an important role in

the pathogenesis of age-related neurodegenerative dis-

eases. The SN and striatum from postmortem PD cases

exhibited significantly greater degrees of oxidative DNA

damage when compared with age-matched control brain

tissue [23]. Even in untreated normal brain, specific

neuroanatomical regions differ in the steady-state levels

of the oxidized DNA base 8-hydroxy-2Vdexoyguanosine(8-oxodG), and such differences are due mainly to brain

region-specific variations in DNA base-excision repair

activity [24]. In addition, the DNA repair response to an

acute oxidative challenge varies across brain regions. The

regions of brain with the least capacity to upregulate 8-

oxoguanine-DNA glycosylase activity (OGG1), mid-

brain and striatum, have been shown to have the highest

levels of oxidative DNA damage following an oxidative

challenge with diethyl maleate [25]. It was not possible

to determine from these studies whether DA neurons are

specifically limited in their capacity to repair oxidative

DNA damage as the data were derived from macro-

dissected brain regions containing heterogeneous popu-

lations of cells, including glia, and various neuronal

phenotypes as well as ependymal and endothelial cells.

In the present report, PC12 cells were used to model a

homogeneous population of neurons on which cell via-

bility, oxidative DNA damage, and DNA repair could be

quantified following toxic exposures. PC12 is a clonal

cell line of rat pheochromocytoma cells that respond to

nerve growth factor (NGF) by extending neurites and

acquiring the appearance of neurons [26]. Both NGF-

treated and untreated cells synthesize, store, secrete, and

take up DA by processes that are similar to those of DA-

ergic neurons [27]. These model DA-ergic cells are

susceptible to manganese cytotoxcity, evidenced by a

report of manganese-induced internucleosomal DNA

fragmentation [4]. PC12 cell cultures incubated with

synthetic melanin also exhibited apoptotic DNA frag-

mentation. [28]. However, the interaction of melanin

with manganese on DNA damage and repair has not

previously been investigated.

The present study was designed to assess the cyto-

toxic effects of melanin and manganese, alone and in

combination, in PC12-derived neurons. Specific objec-

tives were to determine the extent to which cytotoxic

effects of melanin and manganese are mediated by

oxidative DNA damage and to assess whether the ca-

pacity to repair DNA maintains the viability of the

neurons. Understanding the neurotoxic actions of man-

ganese and its interaction with melanin may shed light on

the neurodegenerative process in IPD, but in addition

will be important in elucidating the role of this heavy

metal in the pathophysiology of parkinsonism associated

with occupational exposures to manganese, such as

welding and manganese mining.

For our investigations we have chosen synthetic

melanin produced by oxidative polymerization of DA.

Spectroscopic studies have shown that synthetic melanin

has spectrophotometric absorbance bands identical to,

and exhibits chemical properties similar to those of

natural neuromelanin [29,30]. Synthetic melanin was

found to be an appropriate model for comparative studies

of the influence of natural nigral pigments on the

pathogenesis of DA-ergic neuronal degeneration [28].

MATERIALS AND METHODS

Materials

PC12 cell culture was purchased from ATCC (Man-

assas, VA, USA). Melanin, Sephadex G-75, superoxide

dismutase, xantine oxidase, dithiothreitol, bovine serum

albumin, and acrylamide/bisacrylamide (19/1) mixture

were purchased from Sigma (St. Louis, MO, USA).

V. SAVA et al.1146

Protease inhibitors and 8-oxoguanine DNA glycosylase

(mOGG1) were from Boehringer-Mannheim (Indianap-

olis, IN, USA). Synthetic oligonucleotide containing 8-

oxoG was from Trevigen (Gaithersburg, MD, USA).

[32P]ATP (7000 Ci/mmol) was from ICN Biomedical,

Inc. (Costa Mesa, CA, USA). Phosphorylation buffer, 3V-phosphate free T4 polynucleotide kinase, proteinase K,

nuclease P1, and alkaline phosphatase were from Roche

Diagnostic Company (Indianapolis, IN, USA). mOGG1

antibody was from Alpha Diagnostic (San Antonio, TX,

USA). The ECL Western blotting analysis system was

from Amersham Biosciences (Piscataway, NJ, USA). All

other reagents were ACS grade and from Sigma Chem-

ical Company.

A purified preparation of melanin was employed. The

exclusion of low-molecular-weight contaminants (includ-

ing dopamine) from melanin was carried out on Sepha-

dex G-75 (the column’s dimensions were 1.6 � 40 cm) in

a 50 mM phosphate buffer (pH 7.5) at flow rate of 0.5 ml

min�1, and separation was monitored spectrophotomet-

rically at 280 nm.

Cell culture

PC12 cells were driven into the neuronal phenotype

after pregrowing on RPMI 1640 medium containing 25

mM glucose, 4 mM glutamine, 1 mM sodium pyruvate

and supplemented with 10% heat-inactivated horse se-

rum (JRH Biosciences, Lenexa, KS, USA), 5% fetal calf

serum (HyClone Laboratories, Logan, UT, USA), 25 U/ml

penicillin, and 25 Ag/ml streptomycin. The cells were

cultured at 37jC in a humidified 5% CO2, 95% air

atmosphere in 75 cm2 plates. For differentiation, cells

(60,000 cells/cm2) were plated onto collagen-coated plates

(10 Ag/cm2) in RPMI-1640 supplemented with 100 ng/ml

nerve growth factor (NGF) (Harian Bioproducts, Madi-

son, WI, USA), 25 U/ml penicillin, and 25 Ag/ml strepto-

mycin. Medium was changed every 3 days, and after 4

days of exposure to NGF, cells displayed a differentiated

morphology, and a static population was obtained for an

additional 7 days.

The cytotoxicity experiments were carried out in 24-

well cluster plates coated with collagen. Differentiated

cells were incubated in the presence of MnCl2 and

melanin separately and in combination with different

concentrations. Cells were harvested after 24 and 72

h and viability was assessed using trypan blue.

Assay of intracellular melanin content

PC12-derived neuronal cultures were maintained in

six well cluster plates divided into five groups including

control plate and plates treated with different concen-

trations of Mn (0.1, 1, 10, 100 Ag/ml). Each well

contained 0.7 � 106 cells in 2 ml of medium supple-

mented with different concentrations of Ml (0, 25, 100,

200, 300, 500 Ag/ml). After 24 h incubation with the

toxicants, culture medium was aspirated; cells were

washed in cold PBS three times, followed by centrifuga-

tion until the supernatant was clear. Cells were collected

and lysed by sonication in 1 ml of 10 mM KCl. Cellular

debris was removed by centrifugation of the lysate at

10,000 rpm for 10 min and supernatant was transferred

into a spectrophotometer cuvette. The spectral character-

ization of melanin was performed with a Pharmacia

LKB-Ultrospec III spectrophotometer. The concentration

of Ml was calculated from the calibration curve of

absorbance (at 500 nm) against Ml concentrations.

Measurement of oxidative DNA damage

The steady-state concentration of 8-oxodG was used

as a marker of oxidative DNA damage. The procedure

for DNA isolation was basically the same as reported

before [31]. Briefly, cells were harvested from culture

flasks, sonicated in 10 mM ethylenediaminetetraacetic

acid (EDTA), and centrifuged. The pellet was treated

with DNase-free RNase, followed by digestion with

proteinase K. The protein fraction was separated from

DNA by three consecutive organic extractions. The

DNA was precipitated by ethanol and incubated over-

night at �20jC. The ratio of absorbance at 260 nm to

that at 280 nm was employed for qualification of DNA

purity.

The purified DNA was digested with nuclease P1

following by treatment with alkaline phosphatase. To

prevent errors in the values of 8-oxodG that may arise

due to incomplete hydrolysis [32], DNA was incubated

with sodium acetate at 95jC before treatment with

nuclease P1 and alkaline phosphatase, and gel electro-

phoresis was employed to control the completeness of

digestion and hydrolysis. After complete digestion and

hydrolysis of the DNA, the mixture of deoxynucleo-

sides was analyzed with HPLC using 5% methanol

dissolved in 100 mM sodium acetate (pH 5.2) as a

mobile phase, and 8-oxo-2V-Deoxyguanosine (8-oxodG)

was detected with an electrochemical detector (ESA

Coulochem Model 5100A) at +0.4 V. 2V-Dexoyguano-sine (2dG) was detected at 260 nm in the same sample

using a Perkin Elmer 785A Programmable Absorbance

Detector (Perkin Elmer, Norwalk, CT, USA) connected

in series with the electrochemical detector. The 8-

oxodG level was expressed as the ratio 8-oxodG/2-

dG. Data were recorded, stored, and analyzed on a PC

Pentium computer using ESA 500 Chromatography

Data System Software.

Measurement of 8-oxoguanine-DNA glycosylase activity

(OGG1)

The extraction of OGG1 for enzymatic assay was

performed as described previously [31], but modified

Fig. 1. Effect of melanin and manganese (MnCl2) on viability ofdifferentiated PC12 cells after 72 h of combined exposure. Each pointrepresents the mean value of four replications. TC, test concentrationsof melanin (25 Ag/ml) and manganese (1 Ag/ml) chosen for evaluationof effects on DNA damage and repair.

Melanin/manganese effects on DNA damage/repair 1147

for cell cultures. Briefly, cell cultures were harvested

and homogenized in buffer containing 20 mM Tris base

(pH 8.0), 1 mM EDTA, 1 mM DTT, 0.5 mM spermine,

0.5 mM spermidine, 50% glycerol, and protease inhib-

itors. Homogenates were rocked for 30 min after addi-

tion of 0.1 vol 2.5 M KCl and spun at 14,000 rpm for

30 min. The supernatant was aliquoted and specimens

were kept frozen at –70jC until assay. Protein concen-

tration was measured using the bicinchoninic acid assay

[33].

OGG1 activity was measured by incision assay as

previously described [31]. To prepare 32P-labeled du-

plex oligonucleotide, 20 pmol of synthetic probe con-

taining 8-oxodG (Trevigen) was incubated at 37jC with

[32P]ATP and polynucleotide T4 kinase. To separate the

unincorporated free [32P]ATP, the reaction mixtures

were spun through a G-25 spin column. Complemen-

tary oligonucleotides were annealed in 10 mM Tris (pH

7.8), 100 mM KCl, 1 mM EDTA by heating the

samples 5 min at 80jC and gradually cooling at room

temperature.

Incision reaction (20 Al) contained 40 mM Hepes

(pH 7.6), 5 mM EDTA, 1 mM DTT, 75 mM KCl,

purified bovine serum albumin, 100 fmol of [32P]-

labeled duplex oligonucleotide, and protein extract

(30 Ag). The reaction mixture was incubated at 37jCfor 2 h and placed on ice to terminate the reaction.

Twenty microliters of loading buffer containing 90%

formamide, 10 mM NaOH, and blue–orange dye was

added to each sample. After 5 min of heating at 95jCthe samples were resolved in a denaturing 20% poly-

acrylamide gel containing 7 M urea. The gel was

visualized using the Bio-Rad363 Phosphoimager Sys-

tem, and OGG1 incision activity was calculated as the

amount of radioactivity in the band corresponding to the

specific cleavage product over the total radioactivity in

the lane.

Western immunoblotting

The 8-oxoguanine DNA glycosylases extracted from

cell cultures were separated by 12% SDS–PAGE and

transferred onto a nitrocellulose membrane using a Bio-

Rad Semi-Dry Transblot technique. The membranes

were blocked overnight at 4jC in a solution containing

5% dry milk and Tris-buffered saline (TBS) composed

of 200 mM NaCl and 50 mM Tris–HCl (pH 7.4) and

supplemented with 0.04% Tween 20. The membranes

were rinsed in TBS–Tween mixture and incubated

overnight at 4jC with mOGG1 antibody (Alpha Diag-

nostic, TX, USA) using a 1:1000 dilution of 1% dry

milk prepared in TBS–Tween. After being washed (3 �10 min) with TBS–Tween at 4jC, the membranes were

incubated with goat anti-mouse antibody (1:2000 dilu-

tion) conjugated to horseradish peroxidase (Santa Cruz

Biotechnology, CA) for 1 h at room temperature. The

blot was developed with the ECL kit (Amersham

Biosciences).

Lipid peroxidation

Formation of lipid peroxide derivatives was evaluated

by measuring thiobarbituric acid-reactive substances

(TBARS) according to [21]. Briefly, cell cultures were

homogenized in ice-cold 1.15% KCl (w/v); then 0.4 ml

of the homogenates was mixed with 1 ml of 0.375%

thiobarbituric acid, 15% trichloroacetic acid (w/v), 0.25

N HCl, and 6.8 mM butylated hydroxytoluene, placed in

a boiling water bath for 10 min, removed, and allowed to

cool on ice. Following centrifugation at 3000 rpm for 10

min, the absorbance in the supernatants was measured at

532 nm. The amount of TBARS produced was

expressed as nanomoles of TBARS per milligram of

protein using malondialdehyde bis(dimethyl acetal) for

calibration.

Superoxide dismutase assay

Determination of superoxide dismutase (SOD) activ-

ity in cell cultures was based on inhibition of nitrite

formation in the oxidation of hydroxylammonium

with superoxide anion radical [34]. Nitrite formation

was generated in a mixture containing 25 Al xanthine(15 mM), 25 Al hydroxylammonium chloride (10 mM),

250 Al phosphate buffer (65 mM, pH 7.8), 90 Al distilledwater, and 100 Al xanthine oxidase (0.1 U/ml) used as a

Fig. 4. Effects TC of manganese (Mn), melanin (Mn), and theircombination (Co) on SOD activity of PC 12-derived neurons. Datarepresent the means and SEM of five separate preparations assayed induplicate. Open bars represent cells without oxidant exposure (control).Single asterisks indicate significant differences compared with control( p< .05). Double asterisks indicate significant differences comparedwith the combination of Ml and Mn ( p< .05).

Fig. 2. TBARS accumulation during incubation of PC12-derivedneurons incubated with TC of manganese (Mn), melanin (Ml), and theircombination (Co). Data represent the means and SEM of five differentpreparations assayed in duplicate. Open bars represent cells withoutoxidative exposure (control). Single asterisks indicate significance ofthe differences against control ( p< .05). Double asterisks indicatesignificance of the differences compared with the combination of Mland Mn ( p< 0.05).

V. SAVA et al.1148

starter of the reaction. The inhibitory effect of inherent

SOD was assayed at 25jC during 20 min of incubation

with 10 Al of brain tissue extracts. Determination of the

resulting nitrite was performed on the reaction (20 min at

room temperature) with 0.5 ml sulfanilic acid (3.3 mg/

ml) and 0.5 ml a-naphthylamine (1 mg/ml). Optical

absorbance at 530 nm was measured with an Ultrospec

III spectrophotometer (Pharmacia, LKB). The results

were expressed as units of SOD activity calculated per

milligram of protein. The amount of protein in the

samples was determined using the bicinchoninic acid

assay [33].

Statistical analysis

The results are reported as means F SE (or SD as

specified in each graph) of N independent preparations.

The differences between samples were analyzed with

Fig. 3. Level of 8-oxodG as an index of DNA damage in PC12-derivedneurons incubated with TC of melanin (Ml) and manganese (Mn). Datarepresent the means and SEM of five separate preparations. Open barsrepresent cells without oxidant exposure (control). Asterisks indicatesignificant differences against control ( p< .05).

Student’s t test, and a p < .05 was considered statistically

significant.

RESULTS

The neuronal population derived from a clonal cell

line of rat pheochromocytoma cells responded to NGF by

extending neurites and assuming a neuronal phenotype.

Incubations with melanin alone (25 – 500 Ag/ml) resulted

in minimal effects on cell viability (Fig. 1). Incubation

with manganese alone (10 ng/m to 100 Ag/ml) also had a

minimal effect on cell viability. However, the combina-

tion of melanin and manganese decreased viability of

PC12-derived neurons. Incubation with the combination

of melanin and manganese caused a significant (up to

60%) decrease in viability of the PC12-derived neurons in

a dose-dependent manner. The strongest effect of MnCl2

Fig. 5. OGG1 activity in PC12-derived neurons incubated with TC ofmanganese and melanin for 24 and 72 h. Data represent means F SEMof four independent experiments. Single asterisks indicate significanceof the differences against control ( p< .05). Double asterisks indicatesignificant differences compared with combination of Ml and Mn( p< .05).

Fig. 6. Expression of OGG1 protein at 24 h (open bars) and 72 h (filled bars) after treatment of PC12-derived neurons with TC ofmelanin, manganese, and their combination. The Western blot is presented in the upper panel. Results are averaged from analysis ofthree blots and bars represent SEM. The expression levels were calculated using purified OGG1 (20 ng) as a standard.

Fig. 7. Inhibitory effect of melanin on purified OGG1. The enzyme(5 Ag/ml) was incubated with different concentrations of melanin at37jC for 30 min before incision assay. Data represent means F SEMobtained from four tests.

Melanin/manganese effects on DNA damage/repair 1149

was observed in the presence of 500 Ag/ml melanin. Cell

viability was not proportional to manganese concentra-

tion until levels of 100 ng/ml or greater were used. Doses

higher than 100 ng/ml manganese caused sharp decreases

in cell viability.

To study the interaction of melanin with manganese,

concentrations for each agent were chosen that produced

minimal effects on cell viability. This permitted investi-

gation of DNA damage in the early stages of oxidative

stress. The concentrations of MnCl2 and melanin chosen

were 100 ng/ml and 25 Ag/ml, respectively, as depicted

in Fig. 1 by the test concentration (TC) point.

Development of oxidative stress in differentiated

PC12 cells exposed to TC of manganese and melanin

was evaluated using TBARS as an index of lipid perox-

idation. After 24 and 72 h of incubation, the TBARS

level was significantly increased for all treatments in-

cluding separate application of manganese and melanin

as well as their combination (Fig. 2).

Oxidative DNA damage, indicated by levels of 8-

oxodG, was increased after 24 h of melanin exposure, but

not with MnCl2 alone (Fig. 3). At this point, 25 Ag/ml

melanin resulted in a 6-fold increase in the level of

oxidative damage. The damage was slightly diminished

after 72 h of melanin exposure, but the level of 8-oxodG

was still significantly higher than that of the control.

Manganese alone did not produce alterations in 8-oxodG

in the cells. Unfortunately, the effect of combinations of

these toxicants on levels of 8-oxodG could not be

measured because manganese–melanin bound irrevers-

ibly to the extracted DNA, even after extensive purifica-

tion and hydrolysis, causing technical problems in the

HPLC system.

Measures of SOD activity were affected quite differ-

ently by manganese and melanin (Fig. 4). Manganese

incubation increased SOD activity. It was 2.25 fold

higher for the first day of incubation as compared with

control, and increased up to 2.51-fold on the third day. In

contrast, melanin acted as an inhibitor of SOD, decreas-

ing activity of the enzyme to 31 and 27% at 1 and 3 days,

respectively. The combination of both agents was almost

Fig. 8. Correlations between viability of the neurons derived from PC12cells and OGG1 activity modulated by melanin. The effects werecalculated for different combinations of melanin and manganeseconcentrations. The lines represent data obtained by variation of melaninconcentration (0–500 Ag/ml) for fixed concentrations of manganeseof 100 Ag/ml (closed circles), 10 Ag/ml (open circles), and 1 Ag/ml(open squares).

V. SAVA et al.1150

as effective in stimulating SOD activity as was MnCl2alone.

Steady-state levels of 8-oxodG are determined by the

balance between formation of the oxidized base and its

repair by 8-oxoG-DNA glycosylase (OGG1). To measure

OGG1 activity, an incision assay based on a 5V end-labeled 24-mer oligonucleoide containing a single 8-

oxodG at position 10 was employed as previously

described [14]. The formation of the incision product

was dependent on protein concentration for extracts,

indicating that the reaction had not reached saturating

conditions (not shown). OGG1 activity in PC12-derived

neurons treated with manganese and melanin alone as

Fig. 9. Effect of manganese on intracellular uptake of melanin inMeasurements were performed at the following fixed concentrations of(filled squares), 10 Ag/ml (filled rhombuses), 100 Ag/ml (filled trianglpoint represents the mean value of four replications with SEM within

well as with their combination is shown in Fig. 5. There

was a statistically significant increase in OGG1 activity

in cells incubated with melanin at 24 h, followed by a

decrease at 72 h. The activity of OGG1 at 24 h was also

increased when the combination of manganese and

melanin was employed. Interestingly, the amount of

expressed OGG1 protein evaluated by Western blots

was not significantly altered within each time point of

the experiment by either of the toxicants or their combi-

nation (Fig. 6). Hence, the effect of the toxicants was to

alter enzymatic activity of OGG1 and not expression of

the protein. However, it was notable that OGG1 protein

expression under all incubation conditions was decreased

at 72 h as compared with to 24 h.

The inhibitory effect of melanin on OGG1 activity

was observed in an independent experiment. Figure 7

illustrates a decrease in activity of purified OGG1 en-

zyme caused by incubation with different concentrations

of melanin. Similar dose–response profiles can be de-

rived from the plot of cell viability against melanin

concentration (Fig. 1). The correlations between those

two effects were calculated for different conditions

obtained with various concentrations of manganese.

The correlation factors were 0.99914, 0.90434, and

0.89291 at concentrations of manganese of 100, 10,

and 1 Ag/ml, respectively (Fig. 8).

To determine the extent to which OGG1 inhibition

was due to accumulation of melanin within PC12-de-

rived neurons, determination of intracellular melanin was

performed using the spectrophotometric properties of

melanin. Cell cultures were washed three times, followed

by cell lysis and removal of cellular debris. Intracellular

melanin concentrations increased as a function of extra-

differentiated PC12 cells after 24 h of combined exposure.manganese in culture medium: 0.1 Ag/ml (filled circles), 1 Ag/mles). Open circles represent control (no manganese added). Each10% of mean. Errors bars are not shown for clarity.

Melanin/manganese effects on DNA damage/repair 1151

cellular melanin concentration (Fig. 9). The intracellular

melanin content was estimated to be 40.8 pg per cell

when 500 Ag/ml extracellular melanin was added to the

culture medium. This is about 2.8% of the amount of

extracellular melanin. When the highest concentrations

of melanin (500 Ag/ml) and Mn (100 Ag/ml) were both

added to the medium, the intracellular melanin increased

significantly to 61.2 pg/cell (4.3% of the extracellular

amount).

DISCUSSION

Oxidative stress and cytotoxic effects induced by

synthetic melanin and manganese were investigated in

a homogenous population of PC12-derived neurons.

Melanin and manganese employed alone resulted in

minimal decreases in neuronal viability. In contrast, the

combination of melanin and manganese caused signifi-

cant (up to 60%) decreases in viability of PC12-derived

neurons in a dose-dependent manner. Subsequent studies

of the interaction of melanin with manganese to produce

oxidative stress used concentrations for each agent that

resulted in minimal effects on cell viability. This permit-

ted investigation of early stages of toxicity well before

the degenerative process becomes evident as a decrease

in cell viability.

The range of manganese concentrations used in the

cell viability assay (0.1–100 Ag/ml or 0.1–100 ppm)

encompassed and exceeded the range of concentrations

(1–2.4 ppm) shown to be toxic in rat brain [35].

However, the concentrations of melanin used in this

study were based on concentrations employed in earlier

neurotoxicological studies with PC12 cells [28,36].

For a sense of perspective, it should be noted that the

concentration of normal human neuromelanin in SN

pars compacta (SNpc) increases with age. Between the

ages of 50 and 90, the concentrations of neuromelanin

have been reported to be 2.3 to 3.7 Ag/mg wet wt of

SNpc [37]. In PD cases, the neuromelanin levels were

reduced to 1.5 Ag /mg of SNpc. In our present studies,

melanin was added to the cell culture medium at

concentrations ranging from 25 to 500 Ag/ml. We have

shown that the intracellular fraction of melanin in the

PC12-derived neurons reached 2.8% of the extracellular

concentration when 500 Ag/ml was applied, and it

increased to 4.3% in combination with 100 Ag/ml of

MnCl2. However, the test doses of melanin and Mn used

to study effects on oxidative DNA damage and repair

were 25 and 0.1 Ag/ml, respectively, levels that would

result in an intracellular concentration of 13.6 pg mel-

anin per cell (Fig. 9). Assuming that the volume of a

single cell is about 5 � 10�6 Al, the intracellular

concentration of melanin can be estimated to be around

2.7 Ag/Al (2700 ppm), that is, within the range of human

SNpc neuromelanin concentrations (2300–3700 ppm

[37]).

Both manganese and melanin treatment stimulated

lipid peroxidation in PC12-derived neurons. In an earlier

study, however, manganese failed to increase lipid per-

oxidation, suggesting that cell death may not be initiated

by oxidative stress alone [38]. Our results showed that

manganese, even at 100 ng/ml, caused elevation of

TBARS, without an excessive accumulation of oxidative

DNA damage (8-oxodG). In contrast, melanin itself

increased both lipid peroxidation and levels of 8-oxodG

(Fig. 3).

Incubation with either manganese or melanin alone

resulted in very different effects on SOD activity. Man-

ganese increased SOD activity 2.25 times greater than

control for the first 24 h of incubation. After 72 h of

manganese incubation, SOD activity was increased 2.51

fold. In contrast, melanin significantly inhibited SOD

activity, but the combination of both agents was almost

as effective in stimulating SOD as MnCl2 alone.

Oxidative DNA damage in differentiated cells was

significantly increased following 24 h of melanin expo-

sure, but not after MnCl2. At this time point, 25 Ag/ml

melanin resulted in a 6-fold increase in the level of

oxidative damage without a significant effect on cell

viability. The levels of 8-oxodG decreased slightly after

72 h of melanin exposure, but remained significantly

higher than control. It was not possible to measure 8-

oxodG levels in cells co-incubated with melanin and Mn

because the extracted DNA contained tightly bound

melanin that could not be clarified during vigorous

DNA purification.

DNA repair, indicated by enzymatic activity of

OGG1, was upregulated by about 20% in response to

a 7-fold increase in 8-oxodG in cells incubated with

single toxicants and their combinations for 24 h. How-

ever, by the third day of incubation, the activity of

OGG1 dropped significantly below control levels. No-

tably, SOD activity remained elevated over control

levels at 72 h after incubation with Mn or the combi-

nation of toxicants, suggesting that a generalized met-

abolic failure cannot be invoked as an explanation for

the diminished OGG1 activity. Interestingly, the actual

amount of OGG1 protein measured on Western blots

was not altered by either of the toxicants or their

combination after 24 h of incubation, suggesting that

the effect of the toxicants was to alter enzymatic

activity of OGG1 and not expression of the protein.

After 72 h, however, there was a decrease in OGG1

protein expression even under control conditions. This

may be due to diminished OGG1 expression and

increased apoptosis in the NGF-treated PC12 cells in

their differentiated state compared with their preneuro-

nal proliferative state [31].

V. SAVA et al.1152

Incubation of pure OGG1 protein with different con-

centrations of melanin for 30 min resulted in inhibition of

the glycosylase activity, with more than 60% inhibition

caused by 500 Ag/ml. The dose–response curve showed

that significant inhibition of OGG1 activity occurred at

relatively low concentrations of melanin. Similar dose–

response profiles illustrate that cell viability is a function

of melanin concentration (Fig. 1). The correlation be-

tween activity of OGG1 and viability (Fig. 8) supports

the hypothesis that OGG1 plays an important role in

maintaining the integrity and health of PC12-derived

neurons.

Previously, we reported [32] that there is difference

between actively dividing and differentiated cells in the

regulation of base-excision repair and DNA damage

accumulation. Such differences may explain the vulner-

ability of postmitotic neurons to oxidative stresses and

neurotoxins. Both melanin and manganese alone and in

combination are agents that enhance generation of free

radicals and their reaction products. Reactive oxygen free

radicals and H2O2 have long been recognized as geno-

toxic agents [39]. The generation of oxygen free radicals

increases with age in mammalian cells and is associated

with various pathophysiological abnormalities including

cancer, cardiovascular diseases, and Parkinson’s and

Alzheimer’s diseases [38,39]. In IPD postmortem brain

tissue, SN and striatum exhibit the highest levels of

oxidative DNA damage [23]. Reaction of oxygen free

radicals with DNA results in production of specific DNA

base lesions and strand breaks [40]. The most critical

mutagenic lesions in mammalian cells can be repaired

primarily by OGG1. Accumulation of 8-oxodG in DNA

extracted from OGG1(�/�) mouse cells and tissues indi-

cates that OGG1 is the major repair enzyme for 8-oxoG

[41]. The extent of oxidative DNA damage in specific

brain regions, indicated by steady-state levels of 8-

oxodG, has been shown to be inversely related to

OGG1 activity in those regions [24]. In addition, the

DNA repair response (OGG1 activity) to an acute oxida-

tive challenge varies across brain regions and is inversely

related to the extent of oxidative DNA damage [25].

The present investigations have disclosed a correla-

tion between viability of PC12-derived neurons and

OGG1 activity modulated by melanin. The inhibitory

effect of melanin on OGG1 activity appears to play a

significant role in the demise of PC12-derived neurons.

However, to assume that suppression of OGG1 activity

alone is the cause for the demise of cells is premature.

Additional studies would need to be designed to deter-

mine whether specific suppression or elimination of

OGG1 expression (and/or activity) would result in cel-

lular degeneration.

The relevance of these experimental findings to the

role of melanin in the demise of DA-ergic neurons in

IPD is of interest. Brain melanin (neuromelanin) has

been considered to have diametrically opposed actions,

either to enhance toxicity or to protect DA-ergic neurons

against toxicants and oxidative stress. It is known that

the more vulnerable ventral tier of dopaminergic neu-

rons in the SN contains less melanin than the more

heavily pigmented and relatively preserved cells of the

dorsal tier [22]. This ventral cell group is the first to

degenerate in IPD. From this perspective, melanin may

confer a neuroprotective advantage to pigment-contain-

ing cells. Such a conclusion was supported by the

observation that heavily melanized neurons were rela-

tively resistant to development of PD [42]. It was

suggested that melanin might directly scavenge free

radical species or decrease oxidative stress by chelation

of transition metals, such as iron. On the other hand,

investigations showed that the melanin-containing cells

were more vulnerable in PD, suggesting that accumula-

tion of melanin in cells indicates an excessive oxidative

stress [43]. From this point of view, accumulation of

melanin within specific neurons might have a deleteri-

ous impact on their viability. Melanin is known to

attract and accumulate endogenous and exogenous

heavy metals [37]. Increased tissue iron, as found in

the pallidum and SN, or other heavy metals such as Mn

might saturate iron-chelating sites on melanin, and a

looser association between iron and melanin might

result in an increased, rather than decreased, production

of free radical species.

Although the role of melanin in brain (neuromelanin)

appears to be ambiguous, it is clear from the present

experiments that the inhibition of oxidative DNA repair

and the viability of neurons are dependent on the

concentration of melanin. Manganese (and perhaps other

heavy metals such as iron) may enhance the incorpora-

tion by healthy cells of extracellular melanin, released by

dying DA-ergic neurons of the SN. Excessive accumu-

lation of neuromelanin will interfere with oxidative DNA

repair and other antioxidative enzymatic processes, lead-

ing to amplification of oxidative stress, the demise of

neurons, and release of their neuromelanin content to be

taken up by healthy cells or ingested by macrophages. It

can be speculated that activated microglia and/or macro-

phages that ingest extracellular melanin are not as vul-

nerable to the effects of OGG1 inhibition and the

sequelae of oxidative stress because they possess the

enhanced DNA repair and antioxidative capacity of

proliferative cells [31].

Although the neuromelanin–heavy metal interaction

may be involved in IPD and in other forms of chemically

induced parkinsonism, it is not clear why manganism in

humans extends beyond the SD to involve striatum and

pallidum that do not bear neuromelanin. One simple

explanation for the vulnerability of these structures is

Melanin/manganese effects on DNA damage/repair 1153

that Mn is most highly concentrated in the striatum and

globus pallidus, at levels sufficient to trigger a crisis of

oxidative stress in the absence of neuromelanin. In the

SN, where Mn is not most highly concentrated following

exposure, its interaction with neuromelanin may play a

significant role in the demise of DA-ergic neurons.

Acknowledgments—Study was supported by a VA Merit Grant andDOD Grant DAMD 17-03-1-0501.

REFERENCES

[1] Cotzias, G. C. Manganese in health and disease. Physiol. Rev.,38:502–532; 1958.

[2] Cook, D. G.; Fahn, S.; Brait, K. A. Chronic manganese intoxica-tion, Arch. Neurol. 30:59–64; 1974.

[3] Calne, D. B.; Chu, N.-S.; Lu, C.-C.; Huang, C.-S.; Olanow, W.Manganism and idiopathic parkinsonism: similarities and differ-ences, Neurology 44:1583–1586; 1994.

[4] Huang, C.-C.; Chu, N.-S.; Lu, C.-S.; Wang, J.-D.; Tsai, J.-L.;Tzeng, J.-L.; Wolters, E. C.; Calne, D. B. Chronic manganeseintoxication, Arch. Neurol. 46:1104–1106; 1989.

[5] Ehringer, H.; Hornykiewicz, O. Verteilung von Noradrenalin undDopamin (3-hydroxytyramin) im Gehirn des Menschen und ihrVerhalten bei Erkrankungen des extrapyramidalen Systems. Klin.Wochenschr. 38:1236–1240; 1960.

[6] Neff, N. H.; Barrett, R. E.; Costa, E. Selective depletion ofcaudate nucleus dopamine and serotonin during chronic manga-nese dioxide administration to squirrel monkeys. Experientia 25:1140–1141; 1969.

[7] Batistatou, A.; Greene, L. A. Aurintricarboxylic acid rescuesPC12 cells and sympathetic neurons from cell death caused bynerve growth factor deprivation: correlation with suppression ofendonuclease activity. J. Cell Biol. 115:461–471; 1991.

[8] Sloot, W. N.; van der Sluijs-Gelling, A. J.; Gramsbergen, J. B. P.Selective lesions by manganese and extensive damage by ironafter injection into rat striatum or hippocampus. J. Neurochem.62:205–216; 1994.

[9] Sanchez-Ramos, J. Toxin-induced parkinsonism, In: Stern, M.;Koller, W. C., eds., Parkinsonian syndromes. New York: RavenPress; 1992:155–171.

[10] Graham, D. G. Catecholamine toxicity: a proposal for the molec-ular pathogenesis of manganese neurotoxicity and Parkinson’sdisease. Neurotoxicology 5:83–95; 1984.

[11] Desole, M. S.; Sciola, L.; Delogu, M. R.; Sircana, S.; Migheli, R.Manganese and 1-methyl-4-(2V-ethylphenyl)-1,2,3,6-tetrahydro-pyridine induce apoptosis in PC12 cells. Neurosci. Lett. 209:193–196; 1996.

[12] Hirata, Y.; Adachi, K.; Kiuchi, K. Activation of JNK pathway andinduction of apoptosis by manganese in PC12 cells. J. Neuro-chem. 71:1607–1615; 1998.

[13] Mochizuki, H.; Nakamura, N.; Nishi, K.; Mizuno, Y. Apoptosis isinduced by 1-methyl-4-phenylpyridinium ion (MPP+) in ventralmesencephalic – striatal coculture in rat. Neurosci. Lett. 170:191–194; 1994.

[14] Tatton, N. A.; Kish, S. J. In situ detection of apoptotic nuclei inthe substantia nigra compacta of 1-methyl-4-phenyl-1,2,3,6-tetra-hydropyridine-treated mice using terminal deoxynucleotidyl trans-ferase labelling and acridine orange staining. Neuroscience 77:1037–1048; 1997.

[15] Malecki, E. A. Manganese toxicity is associated with mitochon-drial dysfunction and DNA fragmentation in rat primary striatalneurons. Brain Res. Bull. 55:225–228; 2001.

[16] Ramsay, R. R.; Singer, T. P. Relation of superoxide generation andlipid peroxidation to the inhibition of NADH-Q oxidoreductase byrotenone, piericidin A, and MPP+. Biochem. Biophys. Res. Com-mun. 189:47–52; 1992.

[17] Vyas, I.; Heikkila, R. E.; Nicklas, W. J. Studies on the neuro-toxicity of MPTP: inhibition of NAD-linked substrate oxidationby its metabolite, MPP+. J. Neurochem. 46:1501–1507; 1986.

[18] Barbeau, A. Manganese and extrapyramidal disorders (a criticalreview and tribute to Dr. George C. Cotzias). Neurotoxicology5:13–35; 1984.

[19] Calne, D. B.; Chu, N. S.; Huang, C. C.; Lu, C. S.; Olanow, W.Manganism and idiopathic parkinsonism: similarities and differ-ences. Neurology 44:1583–1586; 1994.

[20] Mann, D. M.; Yates, P. O. Possible role of neuromelanin in thepathogenesis of Parkinson’s disease. Mech. Ageing Dev. 21:193–203; 1983.

[21] Marsden, C. D. Neuromelanin and Parkinson’s disease. J. Neural.Transm. Suppl. 19:121–141; 1983.

[22] Gibb, W. Melanin, tyrosine hydroxylase, calbindin and substanceP in the human midbrain and substantia nigra in relation to ni-grostriatal projections and differential neuron susceptibility in Par-kinson’s disease. Brain Res. 581:283–291; 1992.

[23] Sanchez-Ramos, J.; Overvik, E.; Ames, B. N. A marker of oxy-radical-mediated DNA damage (oxo8dG) is increased in nigro-striatum of Parkinson’s disease brain. Neurodegeneration (Exp.Neurol.) 3:197–204; 1994.

[24] Cardozo-Pelaez, F.; Brooks, P. J.; Stedeford, T.; Song, S.; Sanchez-Ramos, J. DNA damage, repair, and antioxidant systems in brainregions: a correlative study. Free Radic. Biol. Med. 28:779–785;2000.

[25] Cardozo-Pelaez, F.; Stedeford, T. J.; Brooks, P. J.; Song, S.; San-chez-Ramos, J. R. Effects of diethylmaleate on DNA damage andrepair in the mouse brain. Free Radic. Biol. Med. 33:292–298;2002.

[26] Greene, L. A.; Tischler, A. S. Establishment of a noradrenergicclonal line of rat adrenal pheochromocytoma cells, which re-spond to nerve growth factor. Proc. Natl. Acad. Sci. USA 73:2424–2428; 1976.

[27] Greene, L. A.; Rein, G. Release storage and uptake of catechol-amines by a clonal cell line of nerve growth factor (NGF) respon-sive pheochromocytoma cells. Brain Res. 129:247–263; 1977.

[28] Offen, D.; Ziv, I.; Barzilai, A.; Gorodin, S.; Glater, E.; Hochman,A.; Melamed, E. Dopamine-melanin induced apoptosis in PC12cells: possible implications for the etiology of Parkinson’s disease.Neurochem. Int. 31(2):207–216; 1997.

[29] Aime, S.; Fasano, M.; Bergamasco, B.; Lopiano, L.; Valente, G.Evidence for a glycidic-lipidic matrix in human neuromelanin,potentially responsible for the enhanced iron sequestering abilityof substantia nigra. J. Neurochem. 62:369–371; 1994.

[30] Faraldi, F.; Reyes, M. G.; Alfieri, E.; Levi, A. C. The agingsubstantia nigra: quantitative histologic study. Panminerva Med.36:103–108; 1994.

[31] Stedeford, T.; Cardozo-Pelaez, C.; Nemeth, N.; Song, S.; Harbi-son, R.; Sanchez-Ramos, J. Comparison of base-excision repaircapacity in proliferating and differentiated PC12 cells followingacute challenge with dieldrin. Free Radic. Biol Med. 31:1272–1278; 2001.

[32] Huang, X.; Powell, J.; Mooney, L. A.; Li, C.; Frenkel, K. Impor-tance of complete DNA digestion in minimizing variability of 8-oxo-dG analyses. Free Radic. Biol. Med. 31:1341–1351; 2001.

[33] Smith, P. K.; Krohn, R. I.; Hermanson, G. T.; Mallia, A. K.; Gart-ner, F. H.; Provenzano, M. D.; Fujimoto, E. K.; Goeke, N. M.;Olson, B. J.; Klenk, D. C. Measurement of protein using bicincho-ninic acid. Anal. Biochem. 150:76–85; 1985.

[34] Elstner, E. F.; Heupel, A. Inhibition of nitrite formation fromhydroxylammonium chloride: a simple assay for superoxide dis-mutase. Anal. Biochem. 70:616–620; 1976.

[35] Normandin, L.; Carrier, G.; Gardiner, P. F.; Kennedy, G.; Hazell,A. S.; Mergler, D.; Butterworth, R. F.; Philippe, S.; Zayed, J.Assessment of bioaccumulation, neuropathology, and neurobe-havior following subchronic (90 days) inhalation in Sprague–Dawley rats exposed to manganese phosphate. Toxicol. Appl.Pharmacol. 183:135–145; 2002.

V. SAVA et al.1154

[36] Hirata, Y. Manganese-induced apoptosis in PC12 cells. Neurotox-icol. Teratol. 24:639–653; 2002.

[37] Zecca, L.; Tampellini, D.; Gatti, A.; Crippa, R.; Eisner, M.; Sulzer,S.; Ito, S.; Fariello, R.; Gallorini, M. The neuromelanin of humansubstantia nigra and its interaction with metals. J. Neural Transm.109:663–672; 2002.

[38] Roth, J. A.; Feng, L.; Walowitz, J.; Browne, R. W. Manga-nese-induced rat pheochromocytoma (PC12) cell death is in-dependent of caspase activation. J. Neurosci. Res. 61:162–171;2000.

[39] Smith, G. J.; Grisham, J. W.; Fatteh, M. M. Kinetics of cell deathinduced in 10T1/2 cells by methyl methanesulfonate and the effectsof extracellular calcium on cell death.Am. J. Pathol. 122:488–492;1986.

[40] Breen, A. P.; Murphy, J. A. Reactions of oxyl radicals with DNA.Free Radic. Biol Med. 18:1033–1077; 1995.

[41] Minowa, O.; Arai, T.; Hirano, M.; Monden, Y.; Nakai, S.; Fukuda,M.; Itoh, M.; Takano, H.; Hippou, Y.; Aburatani, H.; Masumura,K.; Nohmi, T.; Nishimura, S.; Noda, T. Mmh/Ogg1 gene inacti-vation results in accumulation of 8-hydroxyguanine in mice. Proc.Natl. Acad. Sci. USA 97:4156–4161; 2000.

[42] Kastner, A.; Hirsch, E.; Lejeune, O.; Javoy-Agid, F.; Rascol, O.;Agid, Y. Is the vulnerability of neurons in the substantia nigra ofpatients with Parkinson’s disease related to their neuromelanincontent? J Neurochem. 59:1080–1089; 1992.

[43] Hirsch, E.; Graybiel, A.; Agid, Y. Melanized dopamine neurons aredifferentially susceptible to degeneration in Parkinson’s disease.Nature 345–348; 1988.