Time Course Assessment of MethylmercuryEffects on C6 Glioma Cells: SubmicromolarConcentrations Induce Oxidative DNADamage and Apoptosis

Silvana Belletti,1 Guido Orlandini,1 Maria Vittoria Vettori,2 Antonio Mutti,2

Jacopo Uggeri,1 Renato Scandroglio,1 Rossella Alinovi,2 and Rita Gatti1*1Department of Experimental Medicine, Histology Section, University of Parma, Parma, Italy2Department of Clinical Medicine, Nephrology and Health Sciences, University of Parma, Parma, Italy

Organic mercury is a well-known neurotoxicant althoughits mechanism of action has not been fully clarified. Inaddition to a direct effect on neurons, much experimentalevidence supports an involvement of the glial compo-nent. We assessed methylmercury hydroxide (MeHgOH)toxicity in a glial model, C6 glioma cells, exposed in the10�5–10�8 M range. The time course of the effects wasstudied by time-lapse confocal microscopy and supple-mented with biochemical data. We have monitored cellviability and proliferation rate, reactive oxygen species(ROS), mitochondrial transmembrane potential, DNA ox-idation, energetic metabolism and modalities of celldeath. The earliest effect was a measurable ROS gener-ation followed by oxidative DNA damage paralleled by apartial mitochondrial depolarization. The effect on cellviability was dose dependent. TUNEL, caspase activityand real-time morphological observation of calcein-loaded cells showed that apoptosis was the only detect-able mode of cell death within this concentration range.N-acetyl-cysteine (NAC) or reduced glutathione (GSH)completely prevent the apoptotic effect of MeHgOH. Thelowest effective MeHgOH concentration was 10�7 M forROS and DNA OH-adducts generation. The effect ofsubmicromolar concentrations of MeHgOH on C6 cellscould be relevant in the developmental neurotoxicitycaused by low dose, long-term exposures, such as thoseof food origin. In addition, we have shown that the sameconcentrations are effective in the induction of DNA ox-idative damage, with further potential pathogeneticimplications. © 2002 Wiley-Liss, Inc.

Key words: methylmercury; glial model; time-lapse con-focal microscopy; oxidative damage; apoptosis

Methylmercury (MeHg) is a highly toxic environ-mental pollutant. Bioaccumulation through the aquaticfood chain can lead to intoxication, particularly in popu-lations with high intake of fish or fish derivatives. MeHgeasily crosses blood–brain and placental barriers, causingirreparable central nervous system (CNS) injuries both inthe adult and in the developing brain (Clarkson, 1993;

Myers et al., 1997). MeHg exerts direct, potent, toxiceffects on neurons but glial cells seem to be an importanttarget. In this regard, it is suggestive that the largest part ofthe brain burden is located in astrocytes, both in humansand in other primates (Charleston et al., 1994). In addi-tion, in vitro experiments suggest that astrocytes play amodulatory role in MeHg-induced neurotoxicity (Aschner,1996). MeHg inhibits glutamate and aspartate uptake inastrocytes (Albrecht et al., 1993; Aschner et al., 1993),leading to increased extracellular concentration of theseamino acids and eventually to a glutamate-mediated exci-totoxic neuronal death (Aschner, 1996; Trotti et al.,1998). It is noteworthy that MeHg does not affect neuronsin the absence of glutamate (Brookes, 1992) or in thepresence of its receptor antagonists (Park et al., 1996;Aschner et al., 2000).

Monnet-Tschudi (1998) has shown that low con-centrations of MeHg (10�10–10�7 M) cause apoptosis ofthe astroglial component in a three-dimensional cell cul-ture system of fetal rat telencephalon at different develop-mental periods. Moreover, MeHg-induced apoptosis hasbeen demonstrated in D384 human astrocytoma cells(Dare et al., 2001). It is difficult to identify the actualmechanism or mechanisms by which MeHg triggers apo-ptosis, however, because multiple effects on a series ofsubcellular targets have been described in different models.Alteration of calcium homeostasis (Hare et al., 1993),microtubule depolymerization (Miura et al., 1999; Cas-toldi et al., 2000), lysosomal damage (Dare et al., 2001)and reactive oxygen species (ROS) generation (Atchisonand Hare, 1994; Yee and Choi, 1996; InSug et al., 1997;

Contract grant sponsor: MURST; Contract grant number: FIL 2000;Contract grant sponsor: European Community; Contract grant number:QLK4-CT99-1356.

*Correspondence to: Rita Gatti, Department of Experimental Medicine,University of Parma, Via Volturno 39, I-43100 Parma, Italy.E-mail: rita.gatti@unipr.it

Received 2 April 2002; Revised 14 June 2002; Accepted 4 July 2002

Published online 13 September 2002 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/jnr.10419

Journal of Neuroscience Research 70:703–711 (2002)

© 2002 Wiley-Liss, Inc.

Shenker et al., 1999) are among the most acknowledgedeffects, some of which are associated with apoptosis.

Although the direct oxidative properties of mercuryand its compounds are largely accepted, the actual effectsof ROS generation are still controversial. In this regard, itis noteworthy that in vitro experiments concerning theprotective action of different antioxidants and manipula-tion of the intracellular scavenging attitude have led tocontrasting results (Park et al., 1996; Ou et al., 1999; Dareet al., 2001). No matter how ROS are generated, theyhave several intracellular targets with different pathoge-netic implications. Aside from the effects due to lipidperoxidation and thiol groups reactivity, specific oxidativeDNA damage could be of particular relevance, as it hasbeen assessed in human lymphocytes (Ogura et al., 1996;Lee et al., 1997).

To provide further insights on the mechanism ofaction of MeHg, it is important to establish the time-course of events after intoxication and the relevance of thedose in its generation. Most in vitro studies have focusedon relatively high concentrations although in humans,long-term, low-dose exposure is likely to be an issue seenmore often. The inadequate sensitivity of some methodscould miss effects induced by concentrations consideredharmless thus far.

Time-lapse confocal microscopy allowed us to trackmorpho-functional changes and the time span of thedeath of individual cells after methylmercury hydroxide(MeHgOH) intoxication in the range 10�5–10�8 M inC6 cells, a glial model. Visual information was supple-mented with biochemical data; in particular we have mon-itored ROS generation and subcellular distribution, oxi-dative DNA damage, mitochondrial activity, cell viabilityand modalities of death. Our aim was to characterize, fromthe potential subcellular targets of MeHg, those actuallyinvolved in commitment to cell death, and the differentcell responses to a wide range of concentrations.

MATERIALS AND METHODS

Reagents

MeHgOH was obtained from Alpha Aesar (Karlsruhe,Germany). Dulbecco’s Modified Eagle’s Medium (DMEM),3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyl tetrazolium bromide(MTT), L-buthionine-(S,R) sulfoximine (BSO), 8-hydroxy-2�-deoxyguanosine (8-OHdG), reduced form of glutathione (GSH)and N-acetyl-L-cysteine (NAC) were from Sigma-Aldrich (St.Louis, MO). Colorimetric substrate DEVD-pNA for caspase 3 wasfrom Alexis Biochemicals (San Diego, CA). Fetal bovine serum(FBS) and Hanks’ salt solution were from Seromed Biochrom KG(Berlin, Germany). Calcein-acetoxymethylester (calcein-AM), pro-pidium iodide (PI), chloromethyl-dihydrodichlorofluoresceindiacetate (CM-H2DCFDA), 3,3�-dihexyloxacarbocyanine io-dide (DiOC6[3]) were from Molecular Probes (Eugene, OR).ATP Bioluminescence Assay Kit HSII was from BoehringerMannheim Biochemica (Milan, Italy), FragEL™ DNA Frag-mentation Detection Kit was from Oncogene Research Prod-ucts (Cambridge, MA), and the Nunclon BACC2 DNA Kit wasfrom Amersham (Buckinghamshire, England).

Cell Cultures

The C6 rat glioma cell line (ATCC, Rockville, MD) wasgrown routinely in DMEM with 10% FBS on plastic cultureflasks (Costar Corporation, Cambridge, MA). Cells were main-tained at 37°C in a 5% CO2 humidified incubator and subcul-tured twice a week. Treatments with MeHgOH were carriedout in DMEM with 10% FBS and were always started 24 hr afterseeding. Therefore, wherever cited in the text, the concentra-tion of MeHgOH reflects the amount added to the culture.

Assessment of Cell Viability and Proliferation

MTT assay. The MTT assay method was used to assessmitochondrial injury and cell viability. Intact mitochondria re-duce the yellow dye tetrazolium salt (MTT) into a blueformazan product; the amount of formazan formed, propor-tional to the mitochondrial dehydrogenase activity, was mea-sured at 570 nm in a spectrophotometer for microplate. Cellswere seeded (1 � 105 cells/well) onto 96-well plates and incu-bated for 3, 6, and 9 hr with MeHgOH in the range 10�5 to10�8 M. The assay was carried out according to Mosmann(1983).

In separate experiments, cells were seeded as above, pre-incubated for 24 hr with L-buthionine-(S,R)sulfoximine 1 mM(BSO), a specific �-glutamil-cysteine synthetase inhibitor, andthen incubated for 3, 6, and 9 hr with MeHgOH in the range10�5–10�8 M.

Cell count. The effect of MeHgOH treatment on thegrowth of C6 cells was assessed by direct counting of cellpopulation. Cells were seeded at 4 � 104 cells/cm2 onto 24-wells plates in DMEM � 10% FBS. Cells were treated for 6, 12,24, and 48 hr with MeHgOH in the 10�5–10�8 M range. At theend of the treatments, the number of cells was determined witha Coulter Counter ZM (Coulter Electronics Ltd., England),after trypsinization of the cultures.

Markers of Oxidative Stress

Chloromethyl-dihydrodichlorofluorescein diac-etate (CM-H2DCFDA). Using confocal microscopy (CLSM)it is possible to monitor, in real-time, the variation of intracel-lular oxidative activity in a preselected microscopic field. CM-H2DCFDA diffuses passively into cells where its acetate groupsare cleaved by intracellular esterases, allowing the chloromethylgroups to react with intracellular thiols including glutathione.Oxidation yields a fluorescent adduct that is trapped in the cell.The fluorescence emitted by CM-H2DCFDA is proportional toits interaction with intracellular ROS.

Cells were seeded on coverslides at 4 � 104 cells/cm2.Observations were made using a Molecular Dynamics Multi-probe 2001 system (Sunnyvale, CA) equipped with an argonlaser and based on a Nikon inverted microscope. We employeda 60� 1.4NA planapo lens and a 50 �m pinhole. Coverslideswere collected from the incubator and immediately placed in aspecial flow chamber and maintained at 37°C, 5% CO2 in air,throughout image acquisition (Orlandini et al., 1999). Cellswere loaded with 10 �M of CM-H2DCFDA for 50 min, thetime necessary to obtain a leveled signal. Images of a preselectedfield were acquired at baseline, 5, 15, 30, and 60 min after thestart of the treatment and at 60-min intervals thereafter. For thequantitative evaluation we computed the average fluorescenceintensity of the pixels pertaining to cells in each image. The

704 Belletti et al.

measures were carried out, exploiting a specific module of theconfocal microscope software (Image Space software, MolecularDynamics).

Images were composed digitally and printed directly onphotographic paper. Signal intensity was rendered in discretepseudocolor scale (range 0–255) and the relevant palette isincluded in Figure 3. Experiments were carried out in quadru-plicate for each tested concentration (10�5–10�8 M MeHgOH).

The efficiency of the probe was checked by the addition ofa strong oxidant (3 mM hydrogen peroxide) to the culture mediumof the controls that results in an almost immediate, intense rise ofthe signal that saturates within 1–2 min (not shown).

8-hydroxy-2�-deoxyguanosine (8-OHdG) measure-ment. C6 cells were seeded in 25-cm2 flasks (2 � 106 cells/flask) and incubated for 3, 6, and 9 hr with MeHgOH (range10�5–10�7 M). To measure hydroxy adducts to DNA, cellularmaterials from each flask were collected for 8-OHdG analysis.The isolated and purified DNA samples were digested withnuclease P1 for 30 min at 37°C, then incubated with alkalinephosphatase for 60 min. The amount of 8-OHdG present inDNA was measured by HPLC with electrochemical detector asdescribed previously by Kim and Lee (1997).

Evaluation of Mitochondrial Function

Mitochondrial transmembrane potential (��m).Confocal microscopy allowed a real-time visualization ofchanges in ��m. We employed the same flow chamber andmicroscope parameters described in the section regarding mark-ers of oxidative stress. In this series of experiments, cells wereloaded with 40 nM DiOC6. This fluorescent carbocyanineprobe accumulates selectively in mitochondrial membrane andthe intensity of fluorescence is proportional to the amplitude of��m. Treatment with 10�5M MeHgOH was started after30 min loading, when the signal became steady. Images of apreselected microscopic field were acquired at baseline and at30-min intervals after the start of the treatment. Calcein-AMand PI were added at the end of treatment to assess the presenceof apoptotic or necrotic elements according to the techniquedescribed by Gatti et al. (1998). Experiments were repeated inquadruplicate.

Determination of intracellular ATP. Cells wereseeded (1 � 105 cells/well) in 96-well plates and incubated for2, 3, 6, and 9 hr with MeHgOH (range 10�5–10�7 M). Intra-cellular ATP was assessed according to Marcaida et al. (1997),using the ATP Bioluminescence Assay Kit HSII. Each determi-nation was corrected for protein content and expressed as apercentage of control.

Modalities of Cell Death Induced by MeHgOH

Evaluation of cell death mode by confocal micros-copy. Cells were seeded on coverslides at 4 � 104 cells/cm2

and prepared for CLSM observation, as described previously inMaterials and Methods for markers of oxidative stress. Cell deathwas evaluated by calcein-AM and PI loading (Gatti et al., 1998).After baseline image acquisition, morphological changes afterMeHgOH treatment (range 10�5–10�7 M) were recorded fromthe same field at 30-min intervals for 12 hr.

In separate experiments, cells were seeded as above, pre-incubated for 1 hr with the thiol-based antioxidants, reducedglutathione (GSH 1 mM) or NAC (1 mM), and treated for 12 hr

with 10�5 M MeHgOH in the presence of the relevant anti-oxidant.

For quantitative purposes the number of apoptotic eventsfor 100 cells in 10 random microscopic fields were counted(1,000 cells) by two investigators unaware of cell treatment. Thecounts were carried out immediately before and after 12 hr oftreatment for all concentrations tested, and experiments werecarried out in quadruplicate.

TUNEL assay. DNA strand breakage, generated dur-ing the apoptotic process, was visualized in C6 cells usingFragEL™ DNA Fragmentation Detection Kit. Briefly, cellsgrown onto chamber slides™ (Nunc, Roskilde, Denmark),were treated with MeHgOH (range 10�5–10�7 M) for 3, 6, and9 hr, washed with TBS and fixed in 4% formaldehyde for10 min at room temperature. After that, slides were processedfollowing the manufacturer’s suggested procedure. Biotinylatednucleotides were detected using streptavidin-horseradish perox-idase conjugate and diaminobenzidine (DAB). Counterstainingwith methyl green was carried out to discriminate the morphol-ogy of normal from apoptotic cells.

For quantification of apoptotic cells by light microscopy,the number of dying elements from 100 cells was counted in 10random fields for each concentration and time, by two investi-gators unaware of cell treatment. Experiments were carried outin quadruplicate.

Evaluation of CCP32-like caspase activity. Caspase3-like enzyme activity was monitored by the cleavage of acolorimetric substrate DEVD-pNA. 2 � 106 cells were seededin 25-cm2 flasks. Cultures were treated for 6 and 12 hr for eachtested concentration (range10�5–10�7 M). Cells were then har-vested, washed with PBS and lysed on ice in 25 mM HEPES,pH 7.4, 5 mM MgCl2, 0.2% Chaps, 1 mM EDTA, 5 �g/mlaprotinin, 1 �g/ml leupeptin, 1 �g/ml pepstatin, 1 mM PMSF.Cell lysates were incubated in an equal volume of 2� reactionbuffer containing 50 mM HEPES, 20 mM DTT, 2 mM EDTA,0.2% Chaps, 10% sucrose, 100 mM NaCl and 50 �M of specificsubstrate DEVD-pNA for 60 min at 37°C; finally, the relativeabsorbance at 405 nm was measured. Each determination wascorrected for protein content and the total DEVDase activitywas expressed as percentage of control.

Statistical Analysis

The data were compared by two-way ANOVA. Whenthe null hypothesis was rejected, differences between treated andcontrol cells were checked by a post-hoc Dunnett’s Test (SPSSfor Windows 9.0, SPSS Inc., Chicago, IL). A repeated measureANOVA was adopted for the analysis of changes in CM-H2DCFDA fluorescence.

RESULTS

MeHgOH Effects on Cell Viability andProliferation

MTT assay. Changes in MTT metabolism in-duced by MeHgOH are shown in Figure 1a. Concentra-tions of 10�6 M or greater caused a statistically significant,dose-dependent mitochondrial dysfunction. After the ini-tial 3-hr drop we did not record further decreases, inde-pendently of the concentration tested. The 20% decreasein mitochondrial dehydrogenase activity induced by

Methylmercury Effects on C6 Glioma Cells 705

10�7 M MeHgOH did not reach statistical significance.No effects were seen after exposure to 10�8 M MeHgOH.

Preincubation with 1 mM BSO for 24 hr, a pro-oxidant condition, increased the degree of mitochondrialimpairment 2- and 5-fold at the concentrations 10�6 and10�5 M, respectively (Fig. 1b).

Effects on cell proliferation. The effect of Me-HgOH on cell proliferation was measured after 6, 12, 24,and 48 hr of intoxication at a concentration range of 10�5

–10�8 M (Fig. 2). A concentration of 10�7 M or less didnot affect cell proliferation although an arrest that lasted48 hr was recorded at the concentration of 10�6 M. Fromthis time on we observed a progressive recovery of pro-liferation (not shown). A concentration of 10�5 M provedlethal for the whole culture.

Oxidative Stress AssessmentCM-H2DCFDA. To visualize intracellular oxi-

dative activity by confocal microscopy, cells were

loaded with the fluorescent probe CM-H2DCFDA(Fig. 3a– c). Figure 3a shows the basal loading of theculture yielding a weak and diffuse signal. Treatmentwith 10�5 M MeHgOH caused an almost immediateincrease in fluorescence that was already detectable after5 min, doubled in about 30 min (Fig. 3b), and reacheda plateau (Fig. 3c) in about 2 hr. The distribution anddimensions of cytoplasmic spots indicated a mitochon-drial localization of the probe (arrows), but a compara-ble increase in nuclear fluorescence was also detectable(Fig. 3c).

Figure 4 shows the time course of cell fluorescence atthe three concentrations tested. MeHgOH at concentra-tions of 10�5 and 10�6 M caused a superimposable in-crease in fluorescence (P 0.01, 30 min vs. baseline and2 hr vs. baseline for both concentrations). After treatmentwith 10�7 M MeHg the signal increase was delayed,although it reached a similar top value after 3 hr (P 0.01vs. baseline); however, the outcome was different amongthe concentrations tested. Although the signal was persis-tent in the 10�5 M group, we recorded a decline in 10�6

M and steeper, 10�7 M groups after 4 hr (for both con-centrations, P 0.01 vs. the relevant highest values offluorescence). No changes in fluorescence followed incu-bation with 10�8 M MeHgOH (not shown).

8-hydroxy-2�-deoxyguanosine (8-OHdG) mea-surement. 8-OHdG was measured as a marker of oxi-dative DNA damage (Fig. 5). The generation of this DNAadduct was roughly dose-dependent. At the lowest con-centration (10�7 M) the first significant increase in8-OHdG was recorded after 6 hr, with a significant trendtoward baseline values after 9 hr, although the value wasstill higher compared to the controls. For the 10�6 and10�5 M concentrations, a significant increase in 8-OHdGwas apparent after only 3 hr. A further increase wasevident after 6 hr for both concentrations, with no changeafter 9 hr.

Fig. 1. a: Effects of MeHgOH treatment on viability in C6 cellsassessed by the MTT assay. Cells were treated with increasing concen-trations of MeHgOH (10�8–10�5 M) for 3, 6, and 9 hr as described inMethods. b: The same experiment carried out after 24 hr preincubationwith 1mM BSO. Data are from four separate experiments carried outon four wells for each time and concentration. Results are presented aspercentage of corresponding controls. (mean SD; *P 0.05; **P 0.01; ***P 0.001).

Fig. 2. Growth curves in the presence of different concentrations(10�8–10�5 M) of MeHgOH. Each point is the mean SD of 4measurements, each in quadruplicate.

706 Belletti et al.

Fig. 3. CLSM images of living C6 cells. a–c: Cells loaded withCM-H2DCFDA. a: Untreated culture, baseline signal. b: The samefield after 30 min treatment with 10�5 M MeHgOH. The increase inoxidative activity is diffuse but more marked in mitochondria (arrows)and nuclei. c: Detail of b acquired after 2 hr; signal is further increased.d,e: Living C6 cells loaded with DiOC6. d: In basal condition mito-chondria have a worm-like aspect. e: The same field after 3 hr in thepresence of 10�5 M MeHgOH. The loss of signal points to mitochon-drial depolarization. In cells with preserved mitochondrial polarization,the organelles are clustered in the perinuclear area. At greater magni-fication, mitochondrial swelling is evident (inset). f: The same field

after calcein-AM counterstaining. The high nuclear signal is indicativeof chromatin condensation. The mitochondrial positivity confirms thedissipation of ��m. g–i: Living C6 cells loaded with calcein-AM.g: Baseline probe loading. h: The same field after 6 hr of treatment with10�5 M MeHgOH. Cells have progressed variably through apoptosis.Almost all cells have shrunk; some are spindled-shaped (arrows)whereas others are roundish (arrowheads). i: Detail of the same fieldafter 12 hr of treatment showing apoptotic bodies. The palette includedin a fits to all panels, and represents a discrete scale of 256 intensities,each corresponding to a pseudocolor. Scale bars � 20 �m in a (for a,b),g (for g,h); 10 �m in c,d (for d–f), I; 5 �m in inset.

Methylmercury Effects on C6 Glioma Cells 707

Modifications in Mitochondrial MetabolismMitochondrial transmembrane potential (��m).

Figure 3d–f shows the same microscopic field after expo-sure to 10�5 M MeHgOH. Mitochondria visualized byDiOC6 had a worm-like aspect (Fig. 3d). After 3 hr, mostcells underwent a partial ��m dissipation, as evidenced bythe diffuse signal fading (Fig. 3e). Before this event, mi-tochondria rearranged in a perinuclear clustering withsome elements markedly swollen (Fig. 3e, inset). At thesame time, the sequential loading with calcein-AM (Fig.3f) showed that depolarized mitochondria were brightlyfluorescent whereas the high nuclear signal was due tochromatin condensation, an early feature of apoptosis(Gatti et al., 1998).

Intracellular ATP. Intracellular levels of ATPwere measured after 2, 3, 6, and 9 hr (Fig. 6). Althoughthe 10�7 M concentration was ineffectual, 10�6 M Me-HgOH caused a significant decrease in intracellular ATP,but only after 9 hr. Exposure to 10�5 M was associatedwith decreased ATP levels at all incubation times; how-ever, intracellular ATP never fell below 60% of controlconcentration.

Modalities of Cell DeathApoptosis in living cells. Figure 3g is a confocal

image of calcein-AM loaded, untreated C6 cells. Theprobe allows a good visualization of cell morphology, witha nucleus vs. cytoplasm signal ratio around 3:1. Figure 3hand 3i show the time course of the events occurring in thesame field after treatment with 10�5 M MeHgOH . After6 hr, 60% of the cells had started apoptosis (Fig. 3h) andhad progressed variably through the process. Some ele-ments were spindled and showed an increased nuclearfluorescence due to chromatin condensation (Fig. 3h,arrows), whereas other cells appeared round (arrowheads)and shrunken, with some degree of blebbing. After 12 hr(Fig. 3i), nuclear fragmentation became prominent withevident generation of blebs and apoptotic bodies thatcharacterize terminal phases of programmed cell death. Atthis time and concentration most cells had started theapoptotic process (81.3 8.9%; P 0.01 vs. control, 1.70.5, mean SD)

In cultures treated with 10�6 or 10�7 M MeHgOH(not shown), we found a lower percentage of elementsundergoing the process in the absence of qualitative dif-ferences, as compared to the highest concentration tested.After 12 hr we recorded a percentage of 32.1 3.8(mean SD, P 0.01 vs. control) apoptotic cells forthe 10�6 M concentration. At the same time 10�7 MMeHgOH has addressed to apoptosis 8.7 0.8% of the

Fig. 4. Intracellular CM-H2DCFDA fluorescence variations after Me-HgOH treatment (10�7–10�5 M). P-values and modalities of compu-tation are reported in the relevant sections of the text. Points aremean SD of four separate experiments.

Fig. 5. Effects of increasing concentrations of MeHgOH treatment(10�7–10�5 M) on 8-OHdG levels in C6 cells at 3, 6, and 9 hr. Theresults are plotted as means SD of four separate experiments intriplicate. *P 0.05; **P 0.01.

Fig. 6. Effects of MeHgOH treatment on ATP levels in C6 cells. Cellswere treated with increasing concentrations of MeHgOH (10�7–10�5

M) for 2, 3, 6, and 9 hr. Data are from four separate experiments carriedout on 4 wells for each time and concentration. Results are presentedas percentage of corresponding controls (mean SD, *P 0.05).

708 Belletti et al.

cells (mean SD, statistically not significant vs. control).No propidium iodide-positive cells were detected at anyconcentration or time, thus concentrations up to 10�5 MMeHgOH did not induce necrosis in this cell line.

Treatment with NAC or GSH completely preventedMeHgOH-induced apoptosis at all the concentrationstested at least up to 12 hr from the beginning of intoxi-cation (not shown).

TUNEL. TUNEL assay was adopted for the de-tection of apoptotic nuclei. Figure 7 shows the relevantcontrol (a) and a representative microscopic field of areaction carried out after 9 hr incubation with 10�6 MMeHgOH (b). Nuclei of apoptotic cells were dense andstained intensely (arrowhead). Figure 7c shows the per-centage of apoptotic elements at different times and con-centrations. The effect was roughly dose- and time-dependent. A slight although not statistically significantincrease in apoptotic cells was already detectable at 10�7 M.At the highest concentration, after 9 hr we recorded a peakpercentage of 77 4.1 (mean SD) apoptotic elements.

Caspase 3 activity. Only 10�5 M MeHgOH in-duced a significant increase in caspase 3-like activity. This

increase, after 6 and 12 hr, was 2.5 and 3.5 times therelevant control, respectively (Fig. 8).

DISCUSSIONWe have studied the effects of the exposure to Me-

HgOH in the range 10�5–10�8 M on the C6 cell line, aglial model expressing markers of both oligodendrocytesand astrocytes (Freshney, 2000). By means of time-lapseconfocal microscopy and biochemical data we have out-lined a synopsis of several cell function parameters; thisallowed us to follow the time-sequence of effects andconsequences induced by the intoxication.

The evidence of excitotoxic neuronal death due tofunctional impairment in glutamate and aspartate uptakeby astrocytes suggests a glial involvement in the neuro-toxicant effect of methylmercury (Aschner et al., 2000). Adirect cytotoxic effect has been shown also in several glialmodels (Monnet-Tschudi, 1998; Dare et al., 2001; Sanfe-liu et al., 2001).

The method of analysis resulted critical for the as-sessment of a threshold concentration. 10�5 M MeHgOHproved lethal for the whole culture, whereas 10�6 Mkilled only a part of the cells, in agreement with previousobservations in other glial models (Monnet-Tschudi,1998; Dare et al., 2001). Moreover, growth curvesshowed that this concentration induced an arrest in pro-liferation of the surviving population, which cells wereable to overcome after about 48 hr. This recovery is notdue to a decay of the toxicant because it was not preventedby a daily MeHgOH-supplemented medium renewal.

MTT assay, cell counting, TUNEL and direct con-focal observation of calcein-AM loaded cells did not detecta statistically significant effect on viability for the 10�7 MMeHgOH concentration; however, the latter techniqueshowed a 5-fold excess of apoptosis in the 10�7 M groupas compared to control. This apparent discrepancy couldbe the consequence of the relatively low number of cellsthat can actually be counted. Similar results were obtainedin 3-D model of developing rat brain (Monnet-Tschudi,

Fig. 7. TUNEL assay. a: C6 control cells. b: After 9 hr of treatmentwith MeHgOH 10�6 M; apoptotic nuclei are dark and condensed(arrowhead). c: Concentration-response relationships for apoptosis in-duction after 3, 6, and 9 hr of treatment of C6 cells with increasingconcentrations of MeHgOH (10�7–10�5 M). Bars represent the aver-age of four experiments. Two independent observers counted thepercentage of positive elements out of 1,000 cells (mean SD). *P 0.05; **P 0.01.

Fig. 8. Caspase 3 activation in C6 cells after incubation with increasingconcentrations of MeHgOH (10�7–10�5 M). Results are presented aspercentage of corresponding controls (n � 4; mean SD; *P 0.05).

Methylmercury Effects on C6 Glioma Cells 709

1998) and suggest potential noxious effects after long-term, low-dose exposure.

In our model, the earliest event we recorded afterintoxication with MeHgOH concentration of 10�7 Mand higher was ROS generation as measured by CM-H2DCFDA. For the two highest concentrations, it wasdetectable after 5 min, doubled in 30 min, and reached aplateau in 2 hr. MeHgOH concentration of 10�7 Mcaused a quantitatively similar effect although the genera-tion was more gradual and delayed (3 hr). In culturestreated with 10�5 M MeHgOH, ROS generation per-sisted, whereas at 10�6and 10�7 M a partial but significantdecrease was observed, which was faster at the lowestconcentration.

MeHg could induce an oxidative stress both bydirect and indirect mechanisms of action. As for the latter,the same concentrations of MeHgOH tested in our studyhave been shown to inhibit the uptake of cystine, thelimiting precursor in glutathione synthesis, in astrocytes(Allen et al., 2001). Our biochemical and direct observa-tions are consistent with such an alteration because thetoxicant becomes more effective if an impairment of glu-tathione synthesis (BSO preincubation) is induced; thishypothesis is further supported by the finding that in thispro-oxidant condition we did not record any shift inthreshold concentration. In addition, we have observedthat NAC or GSH completely prevent the noxious effectson cultures by MeHgOH.

The initial rise in CM-H2DCFDA fluorescence tookplace in mitochondria. This is in agreement with studieson isolated rat glial cells that pointed to the electrontransport chain as the main target of MeHg (Yee andChoi, 1996); however, we observed that a slightly delayedbut persistent high nuclear signal was also evident. There-fore, high oxidative activity occurs in two subcellularcompartments with different pathogenetic implications.

The fact that oxidation occurs at the nuclear level isconfirmed by the finding of significant, dose-dependentgeneration of 8-hydroxy-2�-deoxyguanosine, which is ev-ident after 3 hr and increases progressively thereafter at thetwo higher concentrations tested. The amount of DNAadducts generated, and therefore the degree of DNA dam-age, corresponded roughly to the different effects observedfor each order of magnitude in concentration. Although10�5M MeHgOH had a lethal effect, oxidative DNAdamage could be relevant in the arrest in proliferationcaused by the 10�6 M concentration because it can inducegrowth arrest genes (Gadd family), as described in embry-onic rat neurons (Ou et al., 1997). A genotoxic damage byMeHg chloride and Hg chloride have been reported inhuman lymphocytes (Ogura et al., 1996; Lee et al., 1997).MeHgOH at a concentration of 10�7 M caused a biphasicbehavior in 8-hydroxy-2�-deoxyguanosine generation,because after a 6 hr peak, we recorded a significant (P 0.01) trend toward baseline value; this result, togetherwith the preserved proliferative attitude and the limitedcell casualties, could be ascribed to an almost full efficiencyof the DNA reparative processes.

Mitochondria are among the putative targets ofMeHg neurotoxicity (Atchison and Hare, 1994) as well asa fundamental crossroad in the control of cell death (De-sagher and Martinou, 2000; Kroemer and Reed, 2000;Bernardi et al., 2001). Three hours after the start of theintoxication, when ROS generation was clearly on itsway, the decrease in the DiOC6 signal pointed to a partialdissipation of the mitochondrial transmembrane potential(��m), which is usually ascribed to mitochondrial perme-ability transition (MPT) (Bernardi et al., 2001), in agree-ment with observations in astrocytes (Dare et al., 2001)and lymphocytes (Shenker et al., 1999).

In addition, mitochondria rearranged in a perinu-clear cluster, an early feature of apoptosis (Desagher andMartinou, 2000). Our in vivo confocal system allowed usto confirm that MPT had occurred by observation of thesame DiOC6-loaded cells after the addition of calcein,which is usually excluded from mitochondria but movesfrom the cytosol into the matrix after this event (Lemasterset al., 1998). The consequent release of cytochrome c andapoptosis-inducing factor (AIF) would trigger apoptosisvia the caspases cascade. It is accepted that this process doesnot cause a critical mitochondrial dysfunction because thiswould lead to necrosis. Instead, the ATP supply must bespared to allow caspases activation and for apoptosis toproceed normally (Desagher and Martinou, 2000). In ourexperiments, ATP concentration never fell below 60% ofbaseline values for the highest concentration tested. Werecorded also an increase of caspase 3 activity after 6 hralthough only at the highest concentration tested. Theabsence of caspase activation at lower concentrations couldbe due to the inherent low sensitivity of this method,which may not detect the activation of this pathwaydespite the morphological evidence of an abnormal apo-ptotic rate at a concentration of 10�6 M (TUNEL andcalcein). Accordingly, Dare et al. (2001) failed to detectcaspase activation at the same MeHg concentration in anastrocytoma cell line.

Apoptosis was the only mode of cell death observedin this model. Time-lapse confocal microscopy allowed us,using calcein loading, to monitor dynamic structuralchanges and to measure the time span of individual celldeath. After 3 hr the probe shows chromatin condensation(Gatti et al., 1998) and mitochondrial depolarization (Fig3f). All the following steps leading to the formation ofapoptotic bodies were tracked (Fig. 3h–i). This procedureproves particularly useful when only a limited number ofcells is involved and when apoptosis, as in this condition,occurs asynchronously.

We have outlined the sequence of interaction be-tween MeHgOH and C6 cells: the ROS generation isfollowed by concentration-dependent oxidative DNAdamage, as evidenced by the generation of 8-OHdG. Thelatter event is paralleled by a partial dissipation of themitochondrial transmembrane potential due to a perme-ability transition (MPT). A consistent series of eventsleading to apoptosis ensues, including caspase 3 activationwith preserved ATP levels. We have also observed thatNAC or GSH, two thiol-based ROS scavengers, can

710 Belletti et al.

prevent MeHgOH-induced apoptosis in this cell line. Thechemical nature of these antioxidants suggests that theycould exert their action at the beginning of the chain ofevents leading to apoptosis.

Besides the physical loss of glial elements due toapoptosis, our results suggest that different mechanisms(e.g., involving proliferation control or mutation) could beoperating in long-term exposure to low concentrations ofMeHgOH. These results could be relevant to the devel-opmental neurotoxicity of the relatively low doses ofMeHg associated with consumption of contaminated fish.

ACKNOWLEDGMENTSWe thank Dr. S. Cavazzini for her cooperation.

CLSM is a facility of Centro Interfacolta Misure, ParmaUniversity.

REFERENCESAlbrecht J, Talbot M, Kimelberg HK, Aschner M. 1993. The role of

sulfhydryl groups and calcium in the mercuric chloride-induced inhibi-tion of glutamate uptake in rat primary astrocyte cultures. Brain Res607:249–254.

Allen JW, Shanker G, Aschner M. 2001. Methylmercury inhibits the invitro uptake of the glutathione precursor, cystine, in astrocytes, but not inneurons. Brain Res 894:131–140.

Aschner M. 1996. Astrocytes as modulators of mercury-induced neurotox-icity. Neurotoxicology 17:663–669.

Aschner M, Du YL, Gannon M, Kimelberg HK. 1993. Methylmercury-induced alterations in excitatory amino acid transport in rat primaryastrocyte cultures. Brain Res 602:181–186.

Aschner M, Yao CP, Allen JW, Tan KH. 2000. Methylmercury altersglutamate transport in astrocytes. Neurochem Int 37:199–206.

Atchison WD, Hare MF. 1994. Mechanisms of methylmercury-inducedneurotoxicity. FASEB J 8:622–629.

Bernardi P, Petronilli V, Di Lisa F, Forte M. 2001. A mitochondrialperspective on cell death. Trends Biochem Sci 26:112–117.

Brookes N. 1992. In vitro evidence for the role of glutamate in the CNStoxicity of mercury. Toxicology 76:245–256.

Castoldi AF, Barni S, Turin I, Gandini C, Manzo L. 2000. Early acutenecrosis, delayed apoptosis and cytoskeletal breakdown in cultured cere-bellar granule neurons exposed to methylmercury. J Neurosci Res 59:775–787.

Charleston JS, Bolender RP, Mottet NK, Body RL, Vahter ME, BurbacherTM. 1994. Increases in the number of reactive glia in the visual cortex ofMacaca fascicularis following subclinical long-term methyl mercury ex-posure. Toxicol Appl Pharmacol 129:196–206.

Clarkson TW. 1993. Mercury: major issues in environmental health. En-viron Health Perspect 100:31–38.

Dare E, Li W, Zhivotovsky B, Yuan X, Ceccatelli S. 2001. Methylmercuryand H2O2 provoke lysosomal damage in human astrocytoma D384 cellsfollowed by apoptosis. Free Radic Biol Med 30:1347–1356.

Desagher S, Martinou JC. 2000. Mitochondria as the central control pointof apoptosis. Trends Cell Biol 10:369–377.

Freshney R. 2000. Specialized cells. In: Culture of animal cells. New York:Wiley-Liss. p 374–376.

Gatti R, Belletti S, Orlandini G, Bussolati O, Dall’Asta V, Gazzola GC.1998. Comparison of annexin V and calcein-AM as early vital markers ofapoptosis in adherent cells by confocal laser microscopy. J HistochemCytochem 46:895–900.

Hare MF, McGinnis KM, Atchison WD. 1993. Methylmercury increasesintracellular concentrations of Ca�� and heavy metals in NG108-15 cells.J Pharmacol Exp Ther 266:1626–1635.

InSug O, Datar S, Koch CJ, Shapiro IM, Shenker BJ. 1997. Mercuriccompounds inhibit human monocyte function by inducing apoptosis:evidence for formation of reactive oxygen species, development of mi-tochondrial membrane permeability transition and loss of reductive re-serve. Toxicology 124:211–224.

Kim KB, Lee BM. 1997. Oxidative stress to DNA, protein, and antioxidantenzymes (superoxide dismutase and catalase) in rats treated with benzo-(a)pyrene. Cancer Lett 113:205–212.

Kroemer G, Reed JC. 2000. Mitochondrial control of cell death. Nat Med6:513–519.

Lee CH, Lin RH, Liu SH, Lin-Shiau SY. 1997. Distinct genotoxicity ofphenylmercury acetate in human lymphocytes as compared with othermercury compounds. Mutat Res 392:269–276.

Lemasters JJ, Qian T, Elmore SP, Trost LC, Nishimura Y, Herman B,Bradham CA, Brenner DA, Nieminen AL. 1998. Confocal microscopy ofthe mitochondrial permeability transition in necrotic cell killing, apoptosisand autophagy. Biofactors 8:283–285.

Marcaida G, Minana MD, Grisolia S, Felipo V. 1997. Determination ofintracellular ATP in primary cultures of neurons. Brain Res Brain ResProtoc 1:75–78.

Miura K, Koide N, Himeno S, Nakagawa I, Imura N. 1999. The involve-ment of microtubular disruption in methylmercury-induced apoptosis inneuronal and nonneuronal cell lines. Toxicol Appl Pharmacol 160:279–288.

Monnet-Tschudi F. 1998. Induction of apoptosis by mercury compoundsdepends on maturation and is not associated with microglial activation.J Neurosci Res 53:361–367.

Mosmann T. 1983. Rapid colorimetric assay for cellular growth and sur-vival: application to proliferation and cytotoxicity assays. J ImmunolMethods 65:55–63.

Myers GJ, Davidson PW, Shamlaye CF, Axtell CD, Cernichiari E, ChoisyO, Choi A, Cox C, Clarkson TW. 1997. Effects of prenatal methylmer-cury exposure from a high fish diet on developmental milestones in theSeychelles Child Development Study. Neurotoxicology 18:819–829.

Ogura H, Takeuchi T, Morimoto K. 1996. A comparison of the8-hydroxydeoxyguanosine, chromosome aberrations and micronucleustechniques for the assessment of the genotoxicity of mercury compoundsin human blood lymphocytes. Mutat Res 340:175–182.

Orlandini G, Ronda N, Gatti R, Gazzola GC, Borghetti A. 1999.Receptor-ligand internalization. Methods Enzymol 307:340–350.

Ou YC, Thompson SA, Kirchner SC, Kavanagh TJ, Faustman EM. 1997.Induction of growth arrest and DNA damage-inducible genes Gadd45and Gadd153 in primary rodent embryonic cells following exposure tomethylmercury. Toxicol Appl Pharmacol 147:31–38.

Ou YC, White CC, Krejsa CM, Ponce RA, Kavanagh TJ, Faustman EM.1999. The role of intracellular glutathione in methylmercury-inducedtoxicity in embryonic neuronal cells. Neurotoxicology 20:793–804.

Park ST, Lim KT, Chung YT, Kim SU. 1996. Methylmercury-inducedneurotoxicity in cerebral neuron culture is blocked by antioxidants andNMDA receptor antagonists. Neurotoxicology 17:37–45.

Sanfeliu C, Sebastia J, Ki SU. 2001. Methylmercury neurotoxicity incultures of human neurons, astrocytes, neuroblastoma cells. Neurotoxi-cology 22:317–327.

Shenker BJ, Guo TL, O I, Shapiro IM. 1999. Induction of apoptosis inhuman T-cells by methyl mercury: temporal relationship between mito-chondrial dysfunction and loss of reductive reserve. Toxicol Appl Phar-macol 157:23–35.

Trotti D, Danbolt NC, Volterra A. 1998. Glutamate transporters areoxidant-vulnerable: a molecular link between oxidative and excitotoxicneurodegeneration? Trends Pharmacol Sci 19:328–334.

Yee S, Choi BH. 1996. Oxidative stress in neurotoxic effects of methyl-mercury poisoning. Neurotoxicology 17:17–26.

Methylmercury Effects on C6 Glioma Cells 711