Association between variation in the human KCNJ10 potassium ion channel gene and seizure susceptibility.

PURPOSE: Our research program uses genetic linkage and association analysis to identify human seizure sensitivity and resistance alleles. Quantitative trait loci mapping in mice led to identification of genetic variation in the potassium ion channel gene Kcnj10, implicating it as a putative seizure susceptibility gene. The purpose of this work was to translate these animal model data to a human genetic association study. METHODS: We used single stranded conformation polymorphism (SSCP) electrophoresis, DNA sequencing and database searching (NCBI) to identify variation in the human KCNJ10 gene. Restriction fragment length polymorphism (RFLP) analysis, SSCP and Pyrosequencing were used to genotype a single nucleotide polymorphism (SNP, dbSNP rs#1130183) in KCNJ10 in epilepsy patients (n = 407) and unrelated controls (n = 284). The epilepsy group was comprised of patients with refractory mesial temporal lobe epilepsy (n = 153), childhood absence (n = 84), juvenile myoclonic (n = 111) and idiopathic generalized epilepsy not otherwise specified (IGE-NOS, n = 59) and all were of European ancestry. RESULTS: SNP rs#1130183 (C > T) alters amino acid 271 (of 379) from an arginine to a cysteine (R271C). The C allele (Arg) is common with conversion to the T allele (Cys) occurring twice as often in controls compared to epilepsy patients. Contingency analysis documented a statistically significant association between seizure resistance and allele frequency, Mantel-Haenszel chi square = 5.65, d.f. = 1, P = 0.017, odds ratio 0.52, 95% CI 0.33-0.82. CONCLUSION: The T allele of SNP rs#1130183 is associated with seizure resistance when common forms of focal and generalized epilepsy are analyzed as a group. These data suggest that this missense variation in KCNJ10 (or a nearby variation) is related to general seizure susceptibility in humans.

Quantitative genetic study of maximal electroshock seizure threshold in mice: evidence for a major seizure susceptibility locus on distal chromosome 1.

We conducted a quantitative trait locus (QTL) mapping study to dissect the multifactorial nature of maximal electroshock seizure threshold (MEST) in C57BL/6 (B6) and DBA/2 (D2) mice. MEST determination involved a standard paradigm in which 8- to 12-week-old mice received one shock per day with a daily incremental increase in electrical current until a maximal seizure (tonic hindlimb extension) was induced. Mean MEST values in parental strains were separated by over five standard deviation units, with D2 mice showing lower values than B6 mice. The distribution of MEST values in B6xD2 F2 intercrossed mice spanned the entire phenotypic range defined by parental strains. Statistical mapping yielded significant evidence for QTLs on chromosomes 1, 2, 5, and 15, which together explained over 60% of the phenotypic variance in the model. The chromosome 1 QTL represents a locus of major effect, accounting for about one-third of the genetic variance. Experiments involving a congenic strain (B6.D2-Mtv7(a)/Ty) enabled more precise mapping of the chromosome 1 QTL and indicate that it lies in the genetic interval between markers D1Mit145 and D1Mit17. These results support the hypothesis that the distal portion of chromosome 1 harbors a gene(s) that has a fundamental role in regulating seizure susceptibility.

Acute cocaine-induced seizures: differential sensitivity of six inbred mouse strains.

Mature male and female mice from six inbred stains were tested for susceptibility to behavioral seizures induced by a single injection of cocaine. Cocaine was injected ip over a range of doses (50-100 mg/kg) and behavior was monitored for 20 minutes. Seizure end points included latency to forelimb or hindlimb clonus, latency to clonic running seizure and latency to jumping bouncing seizure. A range of strain specific sensitivities was documented with A/J and SJL mice being most sensitive and C57BL/6J most resistant. DBA/2J, BALB/cByJ and NZW/LacJ strains exhibited intermediate sensitivity. EEG recordings were made in SJL, A/J and C57BL/6J mice revealing a close correspondence between electrical activity and behavior. Additionally, levels of cocaine determined in hippocampus and cortex were not different between sensitive and resistant strains. Additional studies of these murine strains may be useful for investigating genetic influences on cocaine-induced seizures.

Cocaine intake by rats correlates with cocaine-induced dopamine changes in the nucleus accumbens shell.

Extracellular dopamine levels were determined by microdialysis in the core and shell of the nucleus accumbens and the frontal cortex of rats before and after an injection of cocaine (20 mg/kg, IP). After removal of the probes, these same animals were then tested for their voluntary intake of cocaine using the two-bottle, free-choice paradigm. Baseline dopamine levels and their responses to an injection of cocaine differed among the three brain areas. No significant correlations were found between baseline dopamine levels in any of the three brain regions and the voluntary cocaine consumption. A significant negative correlation was found between cocaine-induced increases in extracellular dopamine in the shell of the nucleus accumbens and the voluntary intake of cocaine (r = -0.73, p < 0.01). No such correlations were observed in the accumbens core region or the frontal cortex. These results provide further evidence of the role of the accumbal shell region in cocaine preference, and indicate that cocaine-induced increases in dopamine levels play a role in oral cocaine self-administration or preference. In addition, this relatively novel approach in using the same animals for both cocaine induced neurotransmitter responses and cocaine preference studies can also be applied for the study of other neurotransmitters and drugs of abuse.

Mapping loci for pentylenetetrazol-induced seizure susceptibility in mice.

DBA/2J (D2) and C57BL/6J (B6) mice exhibit differential sensitivity to seizures induced by various chemical and physical methods, with D2 mice being relatively sensitive and B6 mice relatively resistant. We conducted studies in mature D2, B6, F1, and F2 intercross mice to investigate behavioral seizure responses to pentylenetetrazol (PTZ) and to map the location of genes that influence this trait. Mice were injected with PTZ and observed for 45 min. Seizure parameters included latencies to focal clonus, generalized clonus, and maximal seizure. Latencies were used to calculate a seizure score that was used for quantitative mapping. F2 mice (n = 511) exhibited a wide range of latencies with two-thirds of the group expressing maximal seizure. Complementary statistical analyses identified loci on proximal (near D1Mit11) and distal chromosome 1 (near D1Mit17) as having the strongest and most significant effects in this model. Another locus of significant effect was detected on chromosome 5 (near D5Mit398). Suggestive evidence for additional PTZ seizure-related loci was detected on chromosomes 3, 4, and 6. Of the seizure-related loci identified in this study, those on chromosomes 1 (distal), 4, and 5 map close to loci previously identified in a similar F2 population tested with kainic acid. Results document that the complex genetic influences controlling seizure response in B6 and D2 mice are partially independent of the nature of the chemoconvulsant stimulus with a locus on distal chromosome 1 being of fundamental importance.

Genotyping microsatellite polymorphisms by agarose gel electrophoresis with ethidium bromide staining: application to quantitative trait loci analysis of seizure susceptibility in mice.

Agarose gel electrophoresis with ethidium bromide staining (AGE/EBS) is an efficient and reliable method for analyzing microsatellite polymorphisms. We report the use of AGE/EBS for analyzing DNA microsatellite polymorphisms in a preliminary quantitative trait loci (QTL) study of seizure susceptibility in which a candidate gene strategy was used to direct initial mapping efforts. F2 intercross progeny, derived from seizure-sensitive DBA/2J (D2) and seizure-resistant C57BL/6J (B6) inbred strains of mice, were tested for their sensitivity to the seizure-inducing effect of pentylenetetrazol (PTZ), a gamma-aminobutyric acid (GABA) receptor antagonist. A semi-automated method is described, in which DNA microsatellites were amplified by polymerase chain reaction (PCR) to yield products of 100-200 base pair (bp) in length. Alleles were separated on 3-6% MetaPhor agarose gels, stained with ethidium bromide, and visualized by ultraviolet (UV) illumination. Univariate analysis of genotype and phenotype data provides evidence for a seizure-related QTL on chromosome 5, near genes coding for the GABAA receptor subunits alpha 5 and gamma 3. Interestingly, this suggestive QTL derives from the more resistant B6 strain, but it nonetheless provides impetus for the characterization of possible strain differences in these two candidate genes. Overall, these results demonstrate that AGE/EBS can be useful for rapid screening of genomic regions of special interest in QTL mapping studies.

Genetic influences on electrical seizure threshold.

C57BL/6J (B6) and DBA/2J (D2) mice have been characterized previously as seizure-resistant and seizure-sensitive, respectively, a distinction based primarily upon a differential response to the convulsant effects of various drugs. In the present study, electroconvulsive shock (ECS) was used to assess maximal electroshock threshold (MET) in B6, D2 and hybrid mice. Results revealed that D2 mice have a significantly lower MET compared to B6 mice. There was also a significant gender effect for B6 and F2 mice with females exhibiting a lower MET compared to males. METs for F1 and F2 intercross mice were intermediate between the two parental strains. The difference in variance between F2 and F1 generation mice indicated that about three-quarters of the total variance is due to genetic influence. Taken together, results of this study suggest that the large difference in MET between B6 and D2 mice is a highly heritable trait which may yield to genetic dissection through use of quantitative trait locus mapping.

Different patterns of truncated prion protein fragments correlate with distinct phenotypes in P102L Gerstmann-Sträussler-Scheinker disease.

The clinicopathological phenotype of the Gerstmann-Sträussler-Scheinker disease (GSS) variant linked to the codon 102 mutation in the prion protein (PrP) gene (GSS P102L) shows a high heterogeneity. This variability also is observed in subjects with the same prion protein gene PRNP haplotype and is independent from the duration of the disease. Immunoblot analysis of brain homogenates from GSS P102L patients showed two major protease-resistant PrP fragments (PrP-res) with molecular masses of approximately 21 and 8 kDa, respectively. The 21-kDa fragment, similar to the PrP-res type 1 described in Creutzfeldt-Jakob disease, was found in five of the seven subjects and correlated with the presence of spongiform degeneration and "synaptic" pattern of PrP deposition whereas the 8-kDa fragment, similar to those described in other variants of GSS, was found in all subjects in brain regions showing PrP-positive multicentric amyloid deposits. These data further indicate that the neuropathology of prion diseases largely depends on the type of PrP-res fragment that forms in vivo. Because the formation of PrP-res fragments of 7-8 kDa with ragged N and C termini is not a feature of Creutzfeldt-Jakob disease or fatal familial insomnia but appears to be shared by most GSS subtypes, it may represent a molecular marker for this disorder.

Kainate and AMPA receptor binding in seizure-prone and seizure-resistant inbred mouse strains.

Glutamate and its receptors represent the major excitatory neurotransmission system in the mammalian brain and are considered important in the pathogenesis of many neurological diseases. The present study describes saturation binding experiments performed to measure the affinity (Kd) and density (Bmax) of kainate and AMPA receptors in striatum, cortex and hippocampus from mature DBA/2J (DBA) and C57BL/6J (C57) mice. Previous studies have documented that these two strains differ significantly in seizure susceptibility, with DBA mice exhibiting greater sensitivity in various convulsant tests compared to C57 mice. Non-linear regression analysis of binding data together with Student's t-test and ANOVA revealed significantly higher densities of kainate receptors in striatum and of AMPA receptors in cortex of DBA mice. C57 mice exhibited higher striatal [3H]AMPA binding. There were no significant differences between the mouse strains in binding sites prepared from hippocampus and no differences in affinity for either receptor in any brain region studied. The results support a role for kainate and AMPA receptors in seizure sensitivity, possibly by influencing glutamate transmission in specific pathways. It is unlikely, however, that these receptors account for the generation of seizures alone but rather cooperate with other glutamatergic and non-glutamatergic neurotransmitter systems.

Mapping murine loci for seizure response to kainic acid.

Mature DBA/2J (D2) mice are very sensitive to seizures induced by various chemical and physical stimuli, whereas C57BL/6J (B6) mice are relatively seizure resistant. We have conducted a genome-wide search for quantitative trait loci (QTLs) influencing the differential sensitivity of these strains to kainic acid (KA)-induced seizures by studying an F2 intercross population. Parental, F1, and F2 mice (8-10 weeks of age) were injected subcutaneously with 25 mg/kg of KA and observed for 3 h. Latencies to focal and generalized seizures and status epilepticus were recorded and used to calculate an overall seizure score. Results of seizure testing indicated that the difference in susceptibility to KA-induced seizures between D2 and B6 mice is a polygenic phenomenon with at least 65% of the variance due to genetic factors. First-pass genome screening (10-cM marker intervals) in F2 progeny (n = 257) documented a QTL of moderate effect on Chromosome (Chr) 1 with a peak LOD score of 5.5 (17% of genetic variance explained) localized between D1Mit30 and D1Mit16. Provisional QTLs of small effect were detected on Chr 11 (D11Mit224-D11Mit14), 15 (D15Mit6-D15Mit46) and 18 (D18Mit9-D18Mit144). Multiple locus models generally confirmed the Mapmaker/QTL results and also provided evidence for another QTL on Chr 4 (D4Mit9). Multilocus analysis of seizure severity suggested that additional loci on Chrs 5 (D5Mit11), 7 (D7Mit66), and 15 (D15Nds2) might also contribute to KA-induced seizure response. Overall, our results document a complex genetic determinism for KA-induced seizures in these mouse strains with contributions from as many as eight QTLs.

The effects of repeated morphine exposure on mu opioid receptor number and affinity in C57BL/6J and DBA/2J mice.

C57BL/6J (B6) mice self-administer substantial quantities of morphine compared to DBA/2J (D2) mice, and most of the genetic component of this strain difference has been attributed to a locus on chromosome 10 in the vicinity of the mu opioid receptor gene. To compare binding characteristics of mu opioid receptor populations between the two strains, mice were given single daily injections of a long-acting preparation of morphine sulfate (80 mg/kg, s.c.) or saline for a period of seven days, and euthanatized six hours after the last injection. Brains were removed and dissected into specific regions. Receptor binding studies were performed on frontal cortex and striatum. Data were analyzed using non-linear regression, and Kd and Bmax comparisons made between strains and treatments. Specific [3H]DAMGO binding in striatum indicates that the density of mu opioid receptors in saline-treated B6 mice and saline-treated D2 mice does not differ significantly. After repeated morphine injection, B6 mice exhibited a decrease in striatal [3H]DAMGO binding, indicating a downregulation of receptor density by approximately 45% (p=.0003 vs saline-treated B6), a phenomenon not observed in D2 mice. In frontal cortex, no differences in [3H]DAMGO binding were observed between strains or treatment groups. These results demonstrate a significant difference between mu opioid receptor regulation in B6 and D2 mice, and may underlie well documented strain differences in specific opioid-related behaviors.

Differential susceptibility to seizures induced by systemic kainic acid treatment in mature DBA/2J and C57BL/6J mice.

Mature DBA/2J (D2) and C57BL/6J (B6) mice aged 9-10 weeks were studied to determine susceptibility to behavioral seizures induced by kainic acid (KA) and the possible influence exerted by differences in metabolism and blood-brain barrier (BBB) transport. Mice were observed for 4 h after subcutaneous (s.c.) KA injection. Behavioral seizure parameters included latency to first seizure (clonus), latency to tonic/clonic seizure, and latency to status epilepticus (SE). At a KA dose of 25 mg/kg, 80% of D2 mice exhibited tonic/clonic seizures, whereas all B6 mice remained seizure-free. At 30 mg/kg, tonic/clonic seizures were observed in 100% of D2 mice and 25% of B6 mice. Of D2 mice exhibiting at least one clonic seizure in response to KA at a dose of 25 mg/kg, 50% entered SE and eventually died. Administration of [3H]KA (6.6 x 10(6) dpm) at doses of 25 mg/kg (convulsive) or 11.1 micrograms (nonconvulsive) to mice of both strains resulted in similar levels of radioactivity in cortex, hippocampus, and cerebellum 30 and 60 min after injection. Bioconversion of [3H]KA to a radiolabeled brain metabolite in vivo could not be documented in mice from either strain. Results confirm previously reported differences between D2 and B6 mice in their relative susceptibility to seizures induced by systemic KA administration and suggest that these differences are not related to strain-specific variation in metabolism or BBB transport of KA. Further studies of these two strains of mice may be useful for investigating genetic influences upon seizure susceptibility.

Rat strain and age differences in kainic acid induced seizures.

This study reports comparative dose-response data for kainic acid (KA) induced seizures in juvenile (35-40 days old) and adult (70-90 days old) Wistar-Furth (WF), Fisher 344 (F344), Sprague-Dawley (SD) and Long-Evans Hooded (LEH) rats. Juvenile male WF (n = 51), F344 (n = 55), SD (n = 60), LEH (n = 50) and adult male WF (n = 48), F344 (n = 52), SD (n = 52), LEH (n = 53) rats were given KA 6, 8, 10, 12 or 14 mg/kg, sc. As previously demonstrated adult WF and F344 rats showed the greatest sensitivity and most reliable convulsant responses to kainic acid; SD and LEH rats were less sensitive and showed more variable convulsant responses. Regardless of strain, all juvenile rats exhibited greater sensitivity and less variable convulsant response to KA compared to adults. This was most evident in juvenile SD and LEH rats. Results suggest that while seizure sensitivity to KA decreases with age, genetic factors may regulate the expression of this resistance.

In vivo modulation of excitatory amino acid receptors: microdialysis studies on N-methyl-D-aspartate-evoked striatal dopamine release and effects of antagonists.

Striatal dopamine (DA) release was measured following intrastriatal (i.s.) administration of N-methyl-D-aspartate (NMDA) to unanesthetized, freely-moving rats. One hour after insertion of a removable microdialysis probe and perfusion with normal Ringer's solution, a modified Ringer's solution containing 100 mM potassium (high-K+ Ringer's) was used to standardize the preparation. DA release following i.s. administration of NMDA (12.5 mM in normal Ringer's) was dose-dependent. When NMDA (12.5 mM) was administered in high-K+ Ringer's, DA release was greatly potentiated. Administration of the competitive NMDA receptor antagonist aminophosphonovalerate (APV) in normal Ringer's prior to treatment with NMDA in high-K+ Ringer's resulted in a significant reduction of DA release compared to control animals. In contrast, administration of APV priot to treatment with NMDA in normal Ringer's resulted in a significantly increased release of DA compared to controls. Administration of the non-competitive NMDA antagonist, dextromethorphan (DXT) prior to treatment with NMDA in normal Ringer's or NMDA in high-K+ Ringer's caused significant reductions of DA release compared to controls. Intrastriatal DXT also caused dose-dependent inhibition of high-K+ Ringer's-induced DA release. Similarly, administration of the non-specific calcium channel blocker, cadmium, prior to treatment with NMDA resulted in a significant decrease when compared to control values. Results of this study indicate that dose-dependent NMDA-induced striatal DA release is greatly potentiated by potassium suggesting that under physiological conditions in vivo, striatal NMDA receptors are mostly inactivated.(ABSTRACT TRUNCATED AT 250 WORDS)

Partial characterization of kainic acid-induced striatal dopamine release using in vivo microdialysis.

The aim of this study was to characterize interactions between striatal kainate (KA) receptors and dopamine (DA) release using in vivo microdialysis. After insertion of a microdialysis probe and establishment of baseline DA release, each preparation was standardized with a pulse of an iso-osmotic solution of 100 mM KCl in Ringer's solution. DA release following pharmacological manipulation was compared to potassium-induced release and expressed as a percent value. In one group of animals, KA (12.5 mM in Ringer's solution) was administered via the microdialysis probe in 2, 3, 5 or 10 min pulses 30 min following standardization with potassium resulting in release of DA which was 15.7 +/- 3.9, 30.3 +/- 11.3, 67.5 +/- 15.0 and 92.9 +/- 19.8% of potassium-induced DA release, respectively. Perfusion of CdCl2 (0.6 mM in Ringer's solution) 30-45 min prior to a 10 min KA pulse significantly reduced KA-induced DA release compared to control values. Intrastriatal administration of kynurenate (Kyn) attenuated KA-induced DA release in a dose-dependent manner. Levels of DA metabolites in striatal perfusates were significantly reduced following KA administration. This effect was partially reversed by cadmium pretreatment but not affected by Kyn pretreatment. Findings of this study indicate that KA induces striatal DA release in a dose-dependent manner, and this effect is at least partially dependent upon activation of calcium channels. Results also indicate dose-dependent inhibition of KA-induced striatal DA release by the excitatory amino acid receptor antagonist, Kyn, suggesting that this compound interacts with striatal KA receptors and that these receptors are involved with modulating striatal DA release in vivo.

Strain differences in convulsive response to the excitotoxin kainic acid.

We describe a strain of rats (Wistar-Furth) that is highly susceptible to the neurotoxic effects of kainic acid (KA) and presents a reliable and quantifiable (with low within-group variability) animal model of status epilepticus. Wistar-Furth rats are more sensitive and demonstrate a less variable convulsant response than Sprague-Dawley and Long-Evans rats when tested for total time in seizure activity, latency to onset of first seizure, latency to status epilepticus, seizure severity scores, and percentage exhibiting behavioral seizures and status epilepticus. Results suggest that significant heterogeneity exists in the rodent population with regard to neuronal sensitivity to an excitotoxic amino acid and indicate that strain differences are an important consideration in studies using KA.

Potentiation of kainic acid epileptogenicity and sparing from neuronal damage by an NMDA receptor antagonist.

The time course and severity of the excitotoxic syndrome induced in rats by s.c. injection of 10 mg/kg kainic acid (KA) was modified by pretreatment with MK801, a non-competitive inhibitor of the NMDA receptor, at doses of 0.1, 1 and 10 mg/kg. A dose-dependent increase in the severity of the KA-induced electrographic (EEG) manifestations of epilepsy was seen after MK801. This consisted of an earlier appearance and higher number of EEG seizures, longer time spent in seizures, and an earlier onset of status epilepticus. In contrast, behavioral seizures were increased only in the 0.1 mg/kg MK801 group, but abolished by higher doses. On the contrary, wet dog shakes were progressively reduced with increasing doses of MK801. Four of the 9 animals receiving KA-only group and 3 of the 10 animals in the 1 and 10 mg MK801 groups were sacrificed 5 days after KA. The brain of the KA-only rats presented diffuse gross and microscopic evidence of hemorrhagic necrosis and neuronal damage; the MK801 rats showed only minimal neuronal loss in the CA3 hippocampal sector. This study demonstrates that neuronal damage and epileptiform activity can be dissociated. Furthermore, it confirms the protective effect of MK801 against neuronal damage caused by multiple factors. Lastly, it emphasizes the need for EEG monitoring in order to accurately assess any epileptic/antiepileptic effect.

Amino acid profiles in Long-Evans rat superior colliculus, visual cortex, and inferior colliculus.

An ultrasensitive triple-column ion-exchange/fluorometric method was utilized to measure the levels of over 30 amino acids and related primary amino compounds in Long-Evans rat superior colliculus (SC), visual cortex (VC) and inferior colliculus (IC). Comparison of levels of amino compounds revealed distinctly different profiles for each region. Major constituents were the neurotransmitters and related compounds glutamate, glutamine, GABA, taurine, aspartate and glycine. Glutathione levels were also relatively high in all three regions. SC exhibited a significantly higher level of GABA and beta-alanine compared to both VC and IC. VC had significantly higher levels of glutamate and taurine. VC exhibited the lowest level of glycine and IC the highest. A time-course experiment using SC documented that levels of eleven of thirty-four compounds, including GABA, were subject to significant postmortem alteration in vitro. SC GABA stability experiments indicated that significant in vitro increases of free GABA levels between 1 and 4 min postmortem were associated with equimolar decreases of conjugated GABA levels.

Morphological and functional studies of rat tectal transplants.

Donor tectal tissue from fetal rats was transplanted to the tectal region of intact host newborn rats. Histological and ultrastructural characteristics of mature tectal grafts were examined with special emphasis placed on glial elements and the interface between transplant and host brain. Most transplants survived, proliferated, and differentiated to mature neurons and glial cells. Mature transplants were found to contain small and medium sized neurons which formed clusters along the transplant border. Neurites traversed the interface between transplant and host brain. There was an absence of phagocytosis, inflammatory reactions and chromatolytic neurons throughout the transplant. Classical staining techniques and GFAP immunoreactive stains revealed only mild focal astrogliosis confined mostly to the subpial and perivascular regions of the transplant-host interface. Ultrastructural examination of the transplant revealed numerous axodendritic synapses with aggregates of synaptic vesicles and well formed postsynaptic membranes. The absence of a leptomeningeal membrane at several points along the transplant-host interface may allow neurites from both sides to establish functional synaptic contacts. Single unit recording combined with anterograde WGA-HRP neuronal tract tracing was used to determine the presence and extent of host retinal innervation of tectal transplants.

Effects of 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) on ultrastructure of nigral neuromelanin in Macaca fascicularis.

In neurons of the substantia nigra (SN) of Macaca fascicularis the administration of parkinsongenic doses of 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) caused morphological changes of the neuromelanic granules. Under light microscopy, the granules appeared more dispersed and larger. Electron microscopy revealed coalescence of granules in large masses with loss of the electrodense component. Phagocytosis of neuromelanin by glial cells was also observed. In several neurons the neuromelanic changes were evident in the presence of morphologically intact mitochondria. These data suggest an interaction between MPTP and neuromelanin that may have relevance to the nigrotropic toxicity of MPTP and are in agreement with observations on neuromelanin in parkinsonian patients.