Oxidative mechanisms underlying methyl mercury neurotoxicity

073657‘wYI $3.oM1+Il.lwl

hr.1.Ded. Nertrmcio~c~. Vol. 9. No. 2.pp. 147-153. IYYI

Pergamon Press plc @ IWI ISDN

Printed in Great Britain.

OXIDATIVE MECHANISMS UNDERLYING METHYL MERCURY NEUROTOXICITY T. SARAFIAN and M. ANTHONY Department

of Pathology (Neuropathology), (Received 13 December

VERITY

UCLA Center for the Health Sciences, 18-170 CHS, Los Angeles, CA 90024-1732, U.S.A.

1989; in revised form 1 February 1990; accepted 12 February 1990)

Abstract-Cerebellar granule cells from 5-12-day-old rats can be incubated in suspensionat 37°C for up to 3 hr with minimal decline in viability. Methyl mercury was found to produce time- and concentrationdependent cell killing with >85% cell death after 3 hr exposure to a concentration of 20 PM. Previously characterized inhibition of protein and RNA synthesis as well as known methyl mercury-induced defects in cellular ATP production have been shown to be incapable of causing this degree of cell death. Here we report that methyl mercury induced a concentration-dependent increase in membrane lipoperoxidation and a rapid decline in reduced glutathione in this cerebellar neuronal preparation. Hydrogen peroxide at 5 mM was found to closely reproduce each of the cytotoxic effects manifested by methyl mercury suggesting that oxidizing conditions produced by methyl mercury may account for the observed cell death. Methyl mercury-induced hpoperoxidation was not the cause of cell death since malonaldehyde production could be blocked by a-tocopherol or EDTA without appreciably protecting against cell death. Significant protection from methyl mercury-induced cell death was observed, with EGTA, deferoxamine and KCN. We propose that oxidative events contribute to the toxic mechanism of action of methyl mercury in isolated cerebellar granule neurons. Key words: methyl mercury, cerebellar granule cell, glutathione levels, lipoperoxidation,

deferoxamine.

Granule neurons of the cerebellum display selective sensitivity in animals chronically exposed to methyl mercury (MeHg).‘“.” Suspensions of rat cerebellar granule cells have provided an excellent model for analysis of this methyl mercury-induced neurotoxicity. Previous studies from this laboratory using this system have revealed strong inhibition of both macromolecule synthesis and ATP generation. ‘H.2”.2’ Further investigation, however, indicated that cell death produced by methyl mercury could not be caused by the measured decreases in macromolecule and ATP synthesis. l9 Thus, alternative mechanisms of cytotoxicity were implicated in these cells. These mechanisms must necessarily be rapidly inducible and fairly selective for the cerebellar granule cell. Several reports indicate that MeHg and other mercurials are capable of producing membrane lipoperoxidation in biological tissues. 26,30Since the brain is known to be exceptionally sensitive to oxidative, free-radical-mediated injury’ and since vitamin E and selenium have been demonstrated to provide some degree of protection against methyl mercury neurotoxicity in vivo,3*6.28a plausible mechanism was suggested for the selective neurotoxicity of MeHg. In this report we describe specific oxidative injuries associated with MeHg-induced cytotoxicity in cerebellar granule cell suspensions.

EXPERIMENTAL

PROCEDURES

Materials

EDTA, EGTA, glutathione, hydrogen peroxide, hypoxanthine, mannitol, pancreatic DNase, o-phthalaldehyde, a-tocopherol succinate, trypsin and xanthine oxidase were purchased from Sigma Chemical Co. Potassium cyanide was from J.T. Baker, methyl mercury chloride from K & K Laboratories and thiobarbituric acid from Calbiochem. Cell preparation and incubation

Sprague-Dawley rat cerebellar granule cell suspensions were prepared from 5-12-day-old pups as described previously. *’ Cell incubations in Krebs-Ringer buffer (10 mM Tris-HCI, pH 7.4, 10 mM NaPO,, 125 mM NaCI, 5 mM KCl, 4 mM MgS04, 10 mM dextrose) were also performed Abbreviation:

MeHg, methyl

mercury.

147

148

‘1. Saralian

and M. A. Verity

as described previously except that 0.01 mg/ml pancreatic DNase was included in in~ubati(~ns to reduce cell clumping. Incubations were initiated by combining aliquots of cell suspensions with Krebs-Ringer buffer containing MeHg and/or other reagents. For experiments involving fltocopherol succinate, cells were preincubated with this compound for 5 min at room temperature prior to exposure to other reagents.

Lipoperoxidation

assa)

Lipoperoxide assays were performed on 200 p,l aliquots of cell suspensions incubated for varying periods at 37°C and quick-frozen in Iiquid nitrogen prior to short term (0.5-3 hr) storage at - 10°C. Assays were performed by the method of Yagi.29 Thawed samples were combined with 100 ~1 of 0.025 M thiobarbituric acid in 50% acetic acid and incubated for 1 hr at 85°C. After 15 min cooling at 4”C, samples were extracted by adding 1 ml of n-butanol, vortexing and centrifuging 1 min at 12,000 g. Malonaldehyde-thiobarbituric acid was measured spectrophotometritally in a Farrand dual wavelength fluorometer with excitation at 515 nm and emission at 553 nm. G~~tat~ione assay

Assays for reduced glutathione were performed by the method of Mokrasch and Teschke’.’ on 400 ~1 aliquots of granule cell suspensions which were centrifuged 30 set at 12,000 g. Cell pellets were extracted in 200 ~10.1 M HzS04 by grinding with a plastic conical pestle. Following 1-2 hr storage at 4°C samples were centrifuged 30 set at 12,~ g and 100 (~1of supernatant was assayed for glutathione content in tubes containing 2.5 ml 0.1 M sodium phosphate buffer, pH 8.0 containing 5 mM EDTA and 0.1 ml 7.5 mM o-phalaldehyde in absolute ethanol. Fluorescence was measured in a Farrand dual wavelength fluorometer with excitation at 350 nm and emission at 420 nm using freshly prepared glutathione as standard. Cell viability

Cell viability was determined using the trypan blue exclusion assay by removing aliquots of cells at indicated times of incubation and counting the number of stained and unstained cells in the presence of 0.1% trypan blue. RESULTS Lipoperoxidase

and GSH levels

In an effort to ascertain the basis for the previously reportedly membrane damage underlying MeHg-induced trypan blue permeability, membrane li~~roxidation was assayed as a function of incubation time in the presence of 0, 10 and 20 FM MeHg (Fig. 1A). MeHg was found to produce time- and concentration-dependent increases in malonaldehyde formation during 37°C incubations. Substantial increases in lipoperoxidation were observed in control cells without commensurate loss of cell viability (Fig. 1B) in marked contrast to cells treated with MeHg. In addition to increased rates of lipoperoxidation, MeHg was found to produce a substantial decline in the intracellular level of reduced glutathione (Fig. 2). Control cells also displayed progressive impairment of GSH levels which declined by 36% over 3 hr. MeHg-induced GSH depletion was concentration and time-dependent and preceded the induced lipoperoxidation and cell death. In fact, loss of GSH could be detected immediately upon addition of MeHg since samples removed at zero time of incubation displayed consistently lower GSH compared to controls. (It should be noted that, typically, 3-5 min were required for processing of these samples which therefore do not represent true zero time.) After 3 hr incubation, 20 PM MeHg produced a 71% decline in GSH relative to control, a 100% increase in lipoperoxidation and a 91% decline in cell viability. Uxidi~in~ agents

The observation of rapid GSH loss in the presence of MeHg suggested the possibility of cellular oxidative injury mediating cell death. To determine if the temporal patterns of lipoperoxidation and cell death could be reproduced by exogenous oxidative agents, granule cells were incubated in the presence of a variety of oxidizing conditions and assayed for GSH content, lipoperoxidation and cell viability (Table 1). Both the free radical generating system, xanthine oxidase-t

MeHg and oxidative

149

neurotoxicity

t 0.5

1.0

3.0

2.0 TIME,

H

Fig. 1. Effect of MeHg on cerebellar granule cell lipoperoxide content (A) and viability (B). Cells were incubated with 0 (a-a), 10 pM (o---o), and 20 uM (0-O) MeHg and aliquots were removed at the times indicated. Assays were performed as described under Experimental Procedures. Values represent means of 10 determinations * S.E.M. Values for lipoperoxidation are expressed as the increase in malonaldehyde content (nmole per mg protein) beginning at time zero when control cells contained 1.05 + 0.13 nmole m~onaldehyde/mg -_protein. P~0.05for all values compared with controls at 1, 2 and 3 hr.

L

0

2

t lncubotlon

Time

3

(h)

Fig. 2. Time-dependent changes in intraceltular [glutathione] expressed as pg GSH/mg protein in the presence of 0 (o-o), 10 pM MeHg (*-a) and 20 PM MeHg (U). Vahres represent means of indicated number of determinations 5 S.E.M. Glutathione was assayed as described under Experimental Procedures. P~0.05compared with controls for all points except for 10 FM MeHg at 2 hr.

Table 1. Oxidative agents and granule cell injury

Control 10 FM MeHg 20 PM MeHg 5 mM H202

Viability

GSH

94+4(10)* 35 + 6 (9)5 9-e3(5) 15*6(5)

WW 48-c8(7) 29 k 3 (6) 40’5(5)

Lipoperoxide (lOON 158r18(7) 199 -e 27 (10) 174k5 (5)

*Values are based on assays performed after 3 hr incubation in the presence of various reagents and represent means t S.E.M. (number of experiments). Viability values are expressed as absolute percentage of trypan blue negative cells. tGlutathione levels were calculated as pg per mg protein and expressed as percentage of control. Control values were 4.1 r 0.29 (9) after 3 hr incubation. $Lipoperoxidation was calculated as nmofes malonaldehyde per mg protein and expressed as a percentage of control. %P
150

‘I‘.Sarafian and M. A. Verity

hypoxanthine and the oxidizing system Fe*+/ascorbate were found to be much less toxic than methyl mercury and much less active in producing lipoperoxidation (data not shown). Hydrogen peroxide at 5 mM was the only agent among those tested which caused high lipoperoxidation accompanied by greatly diminished GSH and viability. cy-Tocopherol

These observations provide evidence of an association between diminished glutathione, increased lipoperoxidation and rapid MeHg-induced cell death. To determine if lipoperoxidation is critical in the genesis of this cell death, studies were performed with a-tocopherol succinate, a known inhibitor of membrane oxidation (and antidote for in vivo MeHg toxicity). Table 2 summarizes these studies and demonstrates that marked inhibition of lipoperoxidation by cw-tocopherol succinate produced only a slight improvement in cell viability after 3 hr exposure to 10 FM MeHg. cu-Tocopherol also greatly diminished the rate of lipoperoxidation in control cells without altering viability. Protective agents

Studies with KCN revealed that when combined with a variety of other agents, this toxin consistently suppressed the rate of lipoperoxidation to levels well below that observed in control cells. Consequently we investigated the effect of KCN on MeHg toxicity. KCN was found to substantially ameliorate all of the cytotoxic manifestations of MeHg (Fig. 3). Three hour lipoperoxide content was reduced by 85% in cells treated with KCN and MeHg compared to MeHg-only treated cells. GSH content was 50% greater in cells treated with both KCN and MeHg Table 2. Effect of a-tocopherol

succinate on MeHg-induced cell injury Malonaldehyde (nmol/mg protein)

Control Control+a-tocopherol, 25 pM MeHg, 10 FM MeHg (10 pM)+a-tocopherol, 25 FM

2.8 zk0.3 (6)’ 1.3+0.4(6) 4.8 -t 0.5 (7)$ 2.6 rf:0.3 (8)§

Viability W) 94.0 + 2 (8)t 96.1 t 2 (8) 37.7 + 8 (7)$ 48.2 k 5 (S)i/

‘Values represent means k S.E.M. (number of experiments). Values for lipoperoxidation are expressed as the increase in malonaldehyde content during 3 hr incubation at 37°C. tviability was determined using the method of trypan blue exclusion following 3 hr incubation at 37°C. $P< 0.01 compared with control using Student’s t test. §PO.O5 compared with MeHg only.

A. Viability

B.

10

GSH Control o----O3mM KCN H

20AM MeHq

-

KCNtMeHq

I

3

C. Lipaperaxide

1

3

Ttme(hi

Fig. 3. Effect of KCN and MeHg on granule cell cytotoxicity. All values are expressed as means of four determinations + S.E.M. A. Viability values are expressed as absolute percentage of viable cells. B. Glutathione values are expressed as fig GSH per mg protein. C. Lipoperoxidation was calculated as the increase in malonaldehyde content after 1 and 3 hr incubation and expressed as A malonaldehyde/mg protein. P
151

McHg and oxidative neurotoxicity

(2.2 + 0.15 Fg/mg protein) compared to MeHg-only cells (1.46 + 19 Fg/mg protein) despite the fact that KCN depleted GSH by 15% in control cells. KCN increased cell viability in MeHgtreated cells from 20% to 52%. In an attempt to verify the involvement of oxidative/free-radical injury in the neurotoxic mechanism of MeHg action, several agents known to interfere with such reactions were examined for the ability to protect against MeHg toxicity in vim (Table 3). Among these agents, only glutathione (GSH) and the specific ion chelator, deferoxamine, proved to be strongly effective in maintaining viability. EDTA and EGTA proved to be marginally effective in protection. Interestingly, the Ca *+ chelator, EGTA, was slightly more effective in this role than was EDTA. The hydroxyl radical scavenger, mannitol, as well as catalase and superoxide dismutase were ineffective in reversing any aspect of MeHg toxicity (data not shown). Both cY-tocopherol and EDTA prevented MeHg-induced lipoperoxidation without substantial improvement in cell viability (Tables 2 and 3). Hyposmotic mediated cell injury (Table 3) was not associated with a stimulation of lipoperoxidation, thereby negating the possibility that MeHg-induced lipoperoxidation was secondary to cell necrosis. Table 3. Effect of chelating Condition 1. 2. 3. 4. 5. 6. 7.

Control Low osmolarity* 20 FM MeHg #3+ I mM EGTA #3+2.5 mM EDTA #3+2.5 mM deferoxamine #3+l mMGSH

agents and GSH Viability 94 t 4 (20); 27%7 (6) 14+s (9) 37*3 (3)s 2927 (6)$ 64 2 12 (S)11 98+2 (4)

on MeHg

cytotoxicity

GSH ( loo) 66k I4 (5) 2024 (9) 32+8 (3) 34+7 (5) 3428 (4)Ei 129k I4 (4)

Lipoperoxidation

(IN X3 + 176? 60 k 62?4 64+6 79+

I6 (7)/l 13 (9) I2 (3)[1 (6) (4) 12 (4)

*Cells were incubated in a solution consisting of 60% double distilled water and 40% Krebs-Ringer buffer. fAssays were performed after 3 hr incubation and values are expressed as indicated in Table I. SP (0. I compared with condition #3 using Student’s I test. #P <0.02 compared with condition #3. If
DISCUSSION We have demonstrated previously that low concentrations of MeHg cause significant toxicity manifested as trypan blue permeability in suspensions of rat cerebellar granule neurons incubated at 37°C for a period of several hours. I’)Although major reductions in rates of macromolecule and ATP synthesis occurred prior to cell death, inhibition of these metabolic pathways by alternative means proved to be insufficiently toxic to account for the magnitude of cell death induced by methyl mercury. Therefore, additional metabolic perturbation by MeHg was mandated. Our previous studies indicated that exposing suspensions of cerebellar granule cells to oxidizing conditions such as H20z resulted in rates and patterns of cell death which were similar to those produced by MeHg. ” However, there was no evidence to indicate that MeHg could generate oxidizing conditions in these cells. Our present studies revealing concentration- and timedependent lipoperoxidation strongly suggests that a potent oxidizing species of some kind is produced by MeHg, which, coupled with intracellular glutathione depletion, would have severe consequences for maintenance of cellular integrity. Although there have been several reports of mercurial-induced oxidative injury in a few tissues,h.‘X no evidence for such damage in brain or other neurologic tissue has previously been presented. Despite clear evidence of MeHg-mediated lipoperoxidation, our data indicate that the measured lipoperoxidation is not the critical target for lethal cell injury. Firstly, control cells displaying >90% viability for 3 hr of incubation displayed > 100% increases in malonaldehyde content when compared to the content of malonaldehyde at zero time of incubation. Secondly, a-tocopherol, a well known plasma membrane-localized free-radical scavenger, completely blocked MeHg-mediated increases in lipoperoxidation without substantially improving cell viability. Thirdly, the divalent cation chelators EDTA and EGTA also effectively blocked control and MeHg-mediated lipoperoxidation without protecting against cell death.

152

f Sarafian ;mcl M. A. C’crir?

Thus, lipoperoxidation is likely not the cause of MeHg-induced cell death. Noncthclehs the observed lipoperoxidation serves as a clear indicator that oxidative events are activated ln tht* presence of MeHg and suggest that biochemical targets other than membrane lipids wwc ;tx critical sites for lethal MeHg injury. It is relevant to note that similar circumstances havr hccn reported in red blood ceils. Davies and Goldberg’ observed that addition of various ox~dtz~ng agents to red blood cells caused both lipid peroxidation and protein de~rad~~tion. However. protein degradation was detectable after 5 min incubation and preceded lipid peroxidation by ,7 hr. Addition of antioxidants to this system decreased lipid peroxidation without affecting protcolysis. Miki et al.” have also reported that free radical initiated oxidant injury to red blood cells produced lipid and protein oxidation and hemolysis. u-Tocopherol suppressed lipid oxidation but could not prevent protein oxidation or hemolysis. Similar conclusions have been reach& 111 studies utilizing isolated hepatocytes exposed to HgCll or thiol-reactive rcagcnts.“.” Use of the thiobarbituric acid assay for measurement of lipid peroxidation may be a suboptgmal approach in certain circumstances due to possible interference by nonspecific agents.’ However. the assay is generally considered to be reliable when used for in virro studies.’ In the present studies the use of appropriate controls and observation of concentrationand time-dependent increases in thiobarbituric acid-reactive substances provide strong evidence for the absence of confounding variables. The possibility that lipoper~~xidation arose as a secondary, n~~nsI~~cihc consequence of cell death was discounted by studies of cells disrupted by hypo-osmolar conditions. These cells displayed no stimulation of thiobarbituric acid-reactive substances (Table 3) Among the various oxidative conditions examined in the present study, H,O- was found to most closely reproduce the patterns of toxic injury observed with MeHg which were rapid GSH depletion, and increased lipoperoxidation associated with prevalent cell death. These findings suggest the possibility that MeHg may act by causing an increased production of HzO, or iome other intermediate generated in the HzOz pathway of cytotoxic injury. Brain mitochondria have been shown to generate Hz02 when electron transport is blocked with antimycin A.” This activity was particularly high in the cerebellum. Cell killing by HzOz in this dose range has been shown to be dependent on metal ion-catalyzed Fenton reactions generating the extremely reactive hydroxyl free-radical species.’ The protective effects of the iron chelator, deferoxamine, against McHginduced lipoperoxidation and cell death in cerebellar granule cells implicates iron-mediated hydroxyl radical formation in the acute toxic mechanism of action of MeHg in these cells. The lack of protection by mannitol, catalase and superoxide dismutase probably reflects their ~naccessibility to the intracellular sites of O,-, H,Oz and .OH generation.‘“-“’ The calcium chelator, EGTA. was also partially protective against MeHg-induced cell killing Although incubations were suggesting that CaZf may play a role in MeHg neurotoxicity. performed in Ca’+-free Krebs-Ringer buffers, it is known that EGTA has the capacity to lower cytoplasmic Ca”+ levels by mass action following extracellular complexation.” Komulainerl and Bondy” have shown that MeHg alters intrasynaptosomal Ca”’ levels by mechanisms distinct from the usual Ca”+ channel functions. Chavez and Holguin’ have shown that Hg”’ can cause release of mitochondrial stores of Ca’+ which would produce detrimental effects on cellular and particularly neuronal systems. Ca’+ has been demonstrated to aggravate oxidative free-radical injury in isolated mitochondria, synaptosomes and cultured spinal cord neurons.’ ” The protective effects of KCN on MeHg cytotoxicity were enigmatic. A potent neurotoxicant in its own right, cyanide was expected to aggravate the injuries caused by MeHg. Instead. KCN blocked lipoperoxidation and resulted in considerable sparing of neurons from MeHg-induced death. Various hypotheses can be presented as speculations on the basis for this protection: (I) KCN inhibits the terminal step in the mitochondrial electron transport chain, i.e. reduction of molecular oxygen by cytochrome oxidase. If MeHg were to act by causing aberrant electron transfer at this step resulting in enhanced superoxide anion production. KCN would block this reaction. (2) KCN may be capable of inhibiting superoxide dismutase which would then prevent conversion of 02- to HzOZ and subsequent hydroxyl radical formation. Bacterial mutants expressing excessive superoxide dismutase activity are known to be more sensitive to oxidant-mediated cytotoxicity.‘” (3) KCN may bind to intracelluIar catalytic iron sites responsible for Fenton-type hydroxyl radical formation. The first hypothesis would be consistent with early observations made in this laboratory revealing MeHg-induced uncoupling of oxidative phosphorylation and electron Such uncoupling is generally associated with accelerated, uncontrolled electron transport.” transport and increased production of partially reduced oxygen species.

MeHg

and

oxidative neurotoxicity

IS3

The present results provide substantive support for the jnvolvement of aberrant oxidative metabolism in the mechanism of acute neurotoxic action of MeHg in cerebeliar granule cell suspensions. It will be very interesting to determine whether or not similar events can be demonstrated in vivo. Acknowledgemenrs-We would like to thank Luiza Vartavarian and Hildegard Huntsman for their excellent technical assistance. and Scott Brooks for clerical services. We would also like to express graditude to Drs Stephen Bondy. Carl LeBel and Louis Chang for their helpful comments and criticisms. The dedicated efforts of Derrick Clevidence. working as an undergraduate summer Fellow. and Benham Badie were also greatly appreciated. This work was supported by NIH grants ES 02573. ES 04722.

REFERENCES BraughlerJ. M., Duncan L. A. and Goodman T. (1985) Calcium enhances in vifro free radical-induced damage to brain synaptosomes, mitochondria and cultured spinal cord neurons. J. Neurochem. 45. 1288-3293. 2. BuegeJ. A. and Aust S. D. (1978) In Methods in Enzymology (eds Fleisher S. and Packer L.). Vol. 52. pp. 302-310. Academic Press. New York. of methyl mercury neurotoxicity by vitamin E. J. 3. Chang L. W., Gilbert M. and Sprecher J. (1978) Modification Environ. Res. 17, 356-366. 4. Chavez E. and Holguin J. A. (198R) Mitochondrial calcium release as induced by Hg’+. J. Bio!. Chem. 263, 3582-3587. 5. Davies K. J. A. and Goldberg A. L. (1987) Oxygen radicals stimulate intracellular proteolysis and lipid peroxidation by independent mechanisms in erythrocytes. J. Biol. Chem. 262, X220-8226. 6. Ganther H. E. (1980) Interactions of vitamin E and selinium with mercury and silver. Ann. N. Y. Acad. Sci. 355, 212-226. 7. Halliwell B. and Gutteridge J. M. C. (l9R.5) Oxygen radicals and the nervous system. Trettds Neurosci. 8, 22-26. 8. Halliwell B. and Grootveld M. (19R7) The measurement of free radical reactions in humans. FEBS Left. 213, Y-14. Ca’+ by neurotoxic organometals: distinc9. Komulainen H. and Bondy S. C. (1987) Increased free intrasynaptosomal tive mechanisms. Toxicol. Appl. Fharmacol. 88, 77-86. 10. Kotyk A. and Kanacek K. (1975) Ceil Membrane Transport. Plenum Press. New York. II. Malis C. D. and Bonventre J. V. (19%) Mechanism of calcium potentiation of oxygen free radical injury to renal mitochondria. A model for post-ischemic and toxic mitochondrial damage. J. Biol. Chem. 261, 14201-14208. 12. Miki M.. Tamai H.. Mino M.. Yamamoto Y. and Niki E. (1987) Free-radical chain oxidation of rat red blood cells by molecular oxygen and its inhibjtion by u-tocopherol. Arch. Biochem. E~ophy.~. 258, 373-380. 13. Mokrasch L. C. and Teschke E. J. (1984) Gtutathione content of cultured cells and rodent brain regions: a specific fluorometric assay. Anal. Biochem. 140, 506-509. 14. Patole M. S.. Swaroop A. and Ramasarma T. (19X6) Generation of H:O, in brain mitochondria. J. Nettrochem. 47, I-X. IS. Reuhl K. R.. Chang L. W. and Townsend J. W. (19X1) Pathological effects of in rcfero methyl mercury exposure on the cerebellum of the golden hamster. 1. Early effects upon the neonatal cerebellar cortex. Environ. Res. 26,2X1-306. 16. Saez J. C.. Kessler J. A., Bennett N. V. L. and Spray D. C. (19X7) Superoxide dismutase protects cultured neurons against death by starvation. Proc. Nar. Acad. Sci. 84, 3056~3059. 17. Sager P. R.. Doherty R. A. and Rodier P. M. (19X2) Effects of methyl mercury on developing mouse cerebcllar cortex. Exp. Nettrol. 77, l7Y-193. IX. Sarafian T. A.. Cheung M. K. and Verity M. A. (19X4) In vitro methyl mercury inhibiti~)n of protein synthesis in neonatal cerebellar perikarya. Ncuroparhot. Appl. Nettrohiol. IO, X5-100. 19 Sarafian T. A.. Hagler J.. Vartavarian L. and Verity M. A. (19X9) Rapid cell death induced hy methyl mercury in suspensions of cerebellar granule neurons. J. Nettropafhol. Exp. Ntwrol. 48, l-10. 20 Sarafian T. A. and Verity M. A. (IYXS) Inhibition of RNA and protein synthesis in isolated cerebcllar cells by in vitro and br viva methyl mercury. Neu~ochem, Purhol. 3, 27-39. 21 Sarafian T. A. and Verity M. A. (19X6) Mechanism of apparent transcription inhibition by methyl mercury in ccrebellar neurons. J. Nettrochtwt. 47, 625-631. 22 Scott M. D.. Meschnick S. R. and Eaton J. W. (1987) Superoxide dismutase-rich bacteria-paradoxical increase in oxidant toxicity. J. Biol. Chcm. 262, 3640-3645. 23. Stacey N. H. and Klaassen C. D. (1981) Inhibition of lipid peroxidation without prevention of cellular injury in isolated rat hepatocytes. Toxicol. Appl. Phurmacol. 58, X-IX. 2-t Staccy N. H. and Kappus H. (1982) Cellular toxicity and lipid peroxidation in response to mercury. To.rico/. Appl. Pht~nnucol. 63, 29-35. 2.5. Starke P. E.. Hoek J. B. and Farber 1. L. (1986) Calcium-dependent and calcium-independent mechanisms ofirreversihlc cell injury in cultured hepatocytcs. J. RioI. Chem. 261, 3(136-3032. 26. Taylor T. J.. Ricdess F. and Kocsis J. J. (1973) The role of Hg’+ and methyl mercury in lipid pcroxidation. Fed. Proc. 32, %I. 27. Verity M. A.. Brown W. J. and Chcung M. (1975) Organic mercurial cnccphalopathy: irr r,ir,o and itt firm effectsof I.

2x. 29. 30.

methyl mercury on synaptosomal respiration. 1. Nettrochem. 25, 759-766. Welch S. D. (197Y) The protective effect of vitamin E and N,N’-diphenyl-p-phenylenediaminc (DDPD) against methyl mercury toxicity in the rat. J. Ntttr. 109. th73-1681. Yagi K. (1976) A simple fluoromctric assay for lipoperoxide in blood plasma. B&hem. Mrd. IS, 212-216. Yonaha M.. Satio M. and Sagai M. (10X3) Stimulation of lipid peroxidation by methyl mercury in rats. Liji# Sci. 32, 1507-1514.