Increased vulnerability of hippocampal neurons to excitotoxic necrosis ...

© 1999 Nature America Inc. • http://medicine.nature.com

ARTICLES

Increased vulnerability of hippocampal neurons to excitotoxic necrosis in presenilin-1 mutant knock-in mice QING GUO1, WEIMING FU1, BRYCE L. SOPHER2, MILES W. MILLER2, CAROL B. WARE3, GEORGE M. MARTIN2 & MARK P. MATTSON1


1

Sanders-Brown Research Center on Aging and Department of Anatomy & Neurobiology, University of Kentucky, Lexington, Kentucky 40536, USA, 2 Department of Pathology and 3Department of Comparative Medicine, University of Washington, Seattle, Washington 98195, USA Correspondence should be addressed to M.P.M.; email: [email protected]

Excitotoxicity, a form of neuronal injury in which excessive activation of glutamate receptors results in cellular calcium overload1,2, has been implicated in the pathogenesis of Alzheimer disease3,4 (AD), although direct evidence is lacking. Mutations in the presenilin-1 (PS1) gene on chromosome 14 are causally linked to many cases of early-onset inherited AD (refs. 5,6). We generated PS1 mutant mice (PS1M146VKI) that express the PS1 M146V targeted allele at normal physiological levels. Although PS1M146VKI mice have no overt mutant phenotype, they are hypersensitive to seizure-induced synaptic degeneration and necrotic neuronal death in the hippocampus. Cultured hippocampal neurons from PS1M146VKI mice have increased vulnerability to death induced by glutamate, which is correlated with perturbed calcium homeostasis, increased oxidative stress and mitochondrial dysfunction. Agents that suppress calcium influx or release and antioxidants protect neurons against the excitotoxic action of the PS1 mutation. These findings establish a direct link between a genetic defect that causes AD and excitotoxic neuronal degeneration, and indicate new avenues for therapeutic intervention in AD patients. To directly address the excitotoxicity hypothesis of AD, we generated PS1 mutant ‘knock-in’ mice in which an exon encoding an AD-linked PS1 mutation was exchanged for the homologous exon in the mouse PS1 gene, resulting in mice that produce (in the homozygous state) only mutant PS1 and no wild-type PS1. PCR primers designed to amplify a 180-bp product from the first coding exon of the murine PS1 gene were used to isolate a genomic DNA clone from a mouse 129/SvJ P1 library. A 1.2-kb SacIHindIII fragment containing exon 5 was subcloned into pAlter-1 and mutagenized with a 39-bp mutagenic oligonucleotide designed to introduce the I45V/M146V double mutation and a BstEII restriction site: Wild-type 5’–ATGATCAGTGTCATTGTCATTATGACCATCCTCCTGGTG–3’

Ile Met BstEII I145V/M146V

5’–ATGATCAGTGTCATTGTCGTGGTGACCATCCTCCTGGTG–3’

Val Val The only amino-acid polymorphism (between mouse and human PS1) in exon 5 of the murine PS1 gene was ‘humanized’ by introduction of the I145V substitution, and also allowed us to introduce a unique restriction enzyme site (BstEII) that faciliNATURE MEDICINE • VOLUME 5 • NUMBER 1 • JANUARY 1999

tated genotyping. The mutagenized DNA and flanking targeting arms were assembled in vectors derived from pZErO-2.1, into which we had inserted additional cloning sites and loxP sites in the appropriate positions. These 5’ and 3’ targeting arms were then subcloned into the pNTK2 targeting vector7, yielding the vector pNTKI (Fig. 1a). The assembled vector was linearized with PvuI and electroporated into 129/Sv-derived R1 embryonic stem cells8. Genomic DNA from the 250 clones that survived double selection (250 µg/ml G418 and 2 µM ganciclovir) was digested with HindIII and BglI and analyzed by Southern blot using a 600-bp HindIII–HpaI 5’ probe. Nineteen clones produced the expected HindIII and BglI polymorphisms (data not shown), and four of these (19, 106, 157 and 179) were injected into recipient blastocysts and transferred to foster mothers to produce male chimeras. These males were mated with C57BL/6 females to produce heterozygous PS1mv(+/-) mice (129/SvXC57BL/6 F1s); data here were obtained from lines 106 and 179. For genotyping, two PCR primers were used to amplify genomic DNA sequences flanking exon 5 before digestion of the amplified DNA with the restriction enzyme BstEII. Expected product sizes for the wild-type and targeted allele after BstEII enzyme digestion are 530 and 350/180 bp, respectively. The mutant PS1 was successfully targeted to the PS1 allele (Fig. 1b). Relative expression levels of mRNA from the wild-type and PS1 mutant allele were determined by RT–PCR; analysis of RNA derived from PS1M146VKI+/- mice indicates that mRNA expression from the targeted PS1 mutant allele is normal (Fig. 1c). These findings are also consistent with northern blot analysis of RNA derived from wild-type and PS1M146VKI(-/-) mice (data not shown). We also evaluated the expression of PS1 protein in these mice by western blot analysis using a rabbit polyclonal antibody that detects full-length PS1 and a C-terminal 18 kDa PS1-derived peptide9,10. Expression levels of full-length PS1 and the 18-kDa PS1 peptide are essentially identical in wild-type, PS1M146VKI(+/-) and PS1M146VKI(-/-) mice (Fig. 1d). The subcellular localization and levels of PS1 protein also seem unaltered in hippocampal pyramidal neurons from PS1M146VKI mice, with PS1 immunoreactivity being present in punctate sites (presumably endoplasmic reticulum) within the cytoplasm of cell bodies and neurites (Fig. 1e). Because the mutant PS1 protein is expressed at normal levels (in the absence of endogenous wildtype mouse PS1), such knock-in mice are expected to provide an animal model that more closely reflects familial AD in humans than do transgenics in which the mutant transgene is overexpressed at supraphysiological levels, and in which wild-type 101


Fig. 1 Generation of PS1M146VKI mice. a, Strategy used to target the M146V FAD mutation to the third coding exon (exon 5) of the murine PS1 gene. The PS1 genomic sequences incorporated into the targeting vector are delimited by the vertical dashed lines. The neo selection cassette (flanked by loxP sites) was inserted into a FseI restriction site positioned 375 bp 5’ of exon 5. The 5’ probe was used to screen embryonic stem cell lines for the HindIII and BglI restriction enzyme polymorphisms introduced by the insertion of the neo selection cassette. Bg, BglI; H, HindIII; Hp, HpaI ; B, BstEII . b, Genotypes of F2 pups generated by intercrossing PS1M146V (+/-) mice derived from embryonic stem cell line 106 (lanes 1–3) or 179 (lanes 4–6). Right margin, sizes are indicated in bp. A 530-bp fragment (containing exon 5) was amplified by PCR and the BstEII polymorphism was demonstrated by the cleavage of the PCR product into two diagnostic bands of 350 and 180 bp. c and d, Expression from the targeted PS1M146V allele. c, Total brain RNA isolated from the mice in b was used to amplify the DNA sequences encoding amino acids 1–298 of PS1 (encoded by exons 3–8). After RT–PCR, the samples were digested with BstEII and separated on a 1.5% TBE gel. The uncut, 868-bp product represents mRNA expressed from the wild-type allele and the digested, 426/442-bp products

represent mRNA expressed from the targeted PS1M146V allele. d, Total brain protein was also isolated and separated by SDS-PAGE (100 µg protein/lane). After transfer of the protein, the membrane was probed with PS1 antibody. e, Confocal laser scanning microscope images of cultured hippocampal neurons from wild-type (left) and PS1M146VKI (right) mice, immunostained with PS1 antibody.

mouse PS1 is also expressed11–13. Homozygous PS1M146VKI mice (up to 16 months of age) have not shown any signs of an overt mutant phenotype, indicating that the targeted M146V mutation does not impair the normal developmental and physiological functions of PS1 (refs. 14,15). These findings are consistent with data showing that ex-

pression of human PS1 protein in mice suppresses the PS1-null phenotype16,17. Because these findings indicated that PS1 mutations result in the gain of an adverse property of the protein, we did a series of experiments to elucidate the nature of this adverse property. Intrahippocampal administration of the excitotoxin kainate to

WT PS-1

a

PS-1 M146V K1

Homo

Hetero

WT

Homo

c

WT

b Hetero

a

530 350 180

1018 510

Homo

Hetero

WT

Homo

49 33

WT

d

Hetero

298

19

e

WT PS-1

c

b

PS-1 M146V K1

WT PS-1

PS-1 M146V K1

Vehicle, 24h

KA, 0.1 µg, 24h

KA, 0.3 µg, 12h

KA, 0.3 µg, 124h

Undamaged neurons/field


DNA marker WT Hetero Homo WT Hetero Homo

ARTICLES

CA1

Fig. 2 Enhanced seizure-induced excitotoxic necrosis in hippocampus of PS1 mutant knock-in mice. a, Cresyl violet-stained sections showing neurons in region CA3 of hippocampus from wild-type (WT PS-1) and PS1M146VKI mice that had been administered either saline (Vehicle) or the indicated doses of kainic acid (KA). There was more extensive and accelerated loss of neurons in the PS1M146VKI mouse than in the wild-type mouse. b, The undamaged pyramidal neurons in the indicated hippocampal subregions were 102

CA3

Hilus

counted in wild-type (WTPS1) and PS1M146VKI mice 12 h after administration of 0.3 µg KA. Values are the mean and s.e.m. of determinations made in six mice. **, P < 0.05; ***, P < 0.01 compared with corresponding value for kainate-treated wild-type mice (ANOVA with Scheffe’s post-hoc tests). c, Electron micrographs showing neuronal cell bodies (upper), synaptic regions (middle), and myelinated axons (lower) in hippocampus 12 h after kainate administration in wild-type (WT PS-1) and PS1M146VKI mice. NATURE MEDICINE • VOLUME 5 • NUMBER 1 • JANUARY 1999


b

WT, vehicle

Neuron survival (% of initial number)

a Neuron survival (% of initial number)

WT

M146VKI

M146VKI, vehicle

WT, glutamate

M146VKI, glutamate

Time (hr)

Glutamate concentration (µM)

d

c PS1 M146V KI

Glutamate

WT PS-1

[Ca2+]c (nM)

Fig. 3 Disruption of cellular calcium homeostasis contributes to the excitotoxicity-enhancing action of mutant PS1. a, Hippocampal cultures were exposed for 24 h to the indicated concentrations of glutamate (horizontal axis), and neuron survival was quantified. 씲, wild-type; 쏆, M146V KI. Values are the mean and s.e.m. of determinations made in four cultures. **, P < 0.01 compared with corresponding wild-type value (ANOVA with Scheffe’s post-hoc tests). b, Cultures were exposed to 50 µM glutamate for the indicated time periods (horizontal axis), and neuron survival was quantified. Values are the mean and s.e.m. of determinations made in four cultures. **, P < 0.01 compared with corresponding wild-type value (ANOVA with Scheffe’s post-hoc tests). c, [Ca2+]C was monitored before and after exposure to 50 µM glutamate. Each line represents the mean [Ca2+]C in 20–30 neurons. Downward arrow, time of glutamate administration. d, [Ca2+]C was measured 1 min before 50 µM glutamate (Basal), at the peak of the glutamate-induced [Ca2+]C rise (Peak), and 5 min later (Plateau). 쏋, wild-type; , PS1M146VKI. Values are the mean and s.e.m. of determinations made in four cultures (20–30 neurons analyzed in each culture). **, P < 0.01; ***, P < 0.001 compared with corresponding value in wild-type neurons (ANOVA with Scheffe’s post-hoc tests).

[Ca2+]c (nM)


ARTICLES

Basal

Peak

Plateau

Time (sec)

adult mice induces severe seizures and consequent degeneration and death of CA3 pyramidal neurons18. Kainate-induced degeneration and death of CA3, CA1 and hilar neurons was significantly accelerated and increased in PS1M146VKI mice compared with that in wild-type mice (Fig. 2a and b). This difference in vulnerability of the hippocampal neurons to excitotoxicity was particularly salient 12 hours after intrahippocampal injection of 0.3 µg kainate. No damage to dentate granule neurons was observed after kainate administration in either wild-type or PS1M146VKI mice (data not shown). Electron microscope examination of cellular ultrastructure in hippocampal tissue from PS1M146VKI mice 12 hours after treatment with 0.3 µg kainate by intrahippocampal injection showed extensive degeneration of CA3 neurons, which manifest typical features of necrosis including swelling and disruption of mitochondria and endoplasmic reticulum, damage to the plasma and nuclear membranes, detachment and aggregation of polyribosomes, and increased presence of lysosomal structures (Fig. 2c). Each of these ultrastructural alterations were much more salient in CA3 neurons of PS1M146V mice than in those of wild-type mice. Examination of the neuropil region in CA1 (in which CA3 neuron axons synapse on CA1 neuron dendrites) showed massive degeneration of both pre- and postsynaptic elements in the PS1M146VKI mice, in contrast to wild-type mice, in which most synaptic terminals remained intact (Fig. 2c). Although no white matter damage was observed in dorsal hippocampal regions 12 hours after administration of 0.3 µg kainate in wild-type mice, extensive damage to myelinated axons was evident in kainate-treated PS1M146VKI mice (Fig. 2c). Given prior studies of excitotoxic hippocampal damage1,2, our data show that hippocampal neurons of mice harboring an AD-linked PS1 mutation are hyper-vulnerable to excitotoxic necrosis. When parallel cultures of dissociated hippocampal cells from wild-type mice and PS1M146VKI mice were exposed to increasNATURE MEDICINE • VOLUME 5 • NUMBER 1 • JANUARY 1999

ing concentrations of glutamate, the neurons from the PS1M146VKI mice showed increased sensitivity to excitotoxicity, as indicated by a leftward shift in the concentration– response curve (Fig. 3a). Time course analysis showed that neurons from PS1M146VKI mice died more rapidly than neurons from wild-type mice; more than 50% of the PS1M146VKI neurons died within 8 hours of exposure to 100 µM glutamate, whereas less than 20% of wild-type neurons were dead at that time point (Fig. 3b). Because calcium overload plays a central part in excitotoxic neuronal death2, we measured cytoplasmic free calcium levels ([Ca2+]C) before and after exposure to glutamate in wild-type and PS1M146VKI hippocampal neurons. Basal [Ca2+]C was essentially identical in neurons from wild-type and PS1M146VKI mice. However, there was a considerable difference in the [Ca2+]C response to glutamate in wild-type and PS1M146VKI neurons. Both the peak and sustained components of the glutamate-induced [Ca2+]C increase were significantly greater in neurons from the PS1M146VKI mice compared with those in neurons from the wild-type mice (Fig. 3c and d). Oxidative stress and mitochondrial dysfunction are early and important contributors to excitotoxic neuronal injury19,20. We measured three different parameters of oxidative stress in wildtype and PS1M146VKI neurons before and after exposure to glutamate. Cellular peroxide levels were assessed using the probe 2,7-dichlorofluorescin diacetate, mitochondrial reactive oxygen species were measured using the probe dihydrorhodamine 123, and levels of membrane lipid peroxidation were quantified using the thiobarbituric reactive substances assay. Glutamate-induced increases of cellular peroxides, mitochondrial reactive oxygen species and membrane lipid peroxidation were each significantly exacerbated in hippocampal neurons from PS1M146VKI mice compared with those in neurons from wild-type mice (Fig. 4a–c). Glutamate-induced mitochondrial membrane depolarization was significantly exacerbated in hippocampal neurons from 103


b DHR123 fluorescence (% of control)

DCF fluorescence (avg. pixel intensity/neuron)

a

Time (min)

c

Rhodamine123 fluorescence (% of control) Time (hr)

PS1M146VKI mice compared with that in neurons from wildtype mice (Fig. 4d). To establish contributions of enhanced calcium responses and oxidative stress to the excitotoxicity-enhancing action of mutant PS1, we used calcium channel antagonists and antioxidants. Both dantrolene (an agent that blocks calcium release from the endoplasmic reticulum) and nifedipine (a blocker of L-type voltage-dependent calcium channels) protected hippocampal neurons from wild-type and PS1M146VKI mice against glutamate toxicity (Fig. 5). Similarly, uric acid (a scavenger of peroxynitrite and hydroxyl radical) and propyl gallate (an antioxidant that scavenges hydrogen peroxide and suppresses membrane lipid peroxidation) provided strong protection against excitotoxic neuronal injury in both wild-type and PS1M146VKI neurons

WT PS-1

104

PS1 M146V KI

Time (hr)

d

TBARS fluorescence (% of control)

Fig. 4 Glutamate-induced oxidative stress is enhanced in hippocampal neurons from PS1 mutant knock-in mice. a–c, Cultures were exposed to 50 µM glutamate for the indicated time periods (horizontal axes), and levels of 2,7-dichlorofluorescein fluorescence (a, a measure of peroxide levels), dihydrorhodamine 123 fluorescence (b, a measure of mitochondrial reactive oxygen species) and thiobarbituric reactive substances fluorescence (c, a measure of membrane lipid peroxidation) were measured in neurons. Values are the mean and s.e.m. of determinations made in four cultures (15–25 neurons analyzed in each culture). a, **, P < 0.01 compared with corresponding value for WTPS1,Vehicle; ***, P < 0.001 compared with corresponding value for M146VKI,Vehicle; b and c, ***, P < 0.001 compared with corresponding value for WTPS1. In a, 10 µM propyl gallate (PG, an antioxidant) or vehicle (0.5% dimethylsulfoxide) were added to cultures 1 h before exposure to glutamate. d, Cultures were exposed for 8 h to the indicated concentrations of glutamate (horizontal axis), and levels of rhodamine 123 fluorescence (an indicator of mitochondrial transmembrane potential) were quantified. Values are the mean and s.e.m. of determinations made in four cultures (20–30 neurons analyzed in each culture). **, P < 0.01 compared with corresponding value for PS1M146VKI mice (ANOVA with Scheffe’s post-hoc tests).

Neuronal survival (% of control)


ARTICLES

Glutamate concentration (µM)

(Fig. 5). These data show that PS1 mutations perturb neuronal calcium homeostasis, and exacerbate oxidative stress and mitochondrial dysfunction, resulting in a lowered threshold for excitotoxic neuronal death. Our data provide the first demonstration, to our knowledge, that a genetic mutation linked to a human neurodegenerative disorder can increase neuronal vulnerability to excitotoxic necrosis, and indicate a role for excitotoxicity in AD. Circumstantial evidence that excitotoxicity plays a part in AD comes from studies showing that neurons that are selectively vulnerable in AD (for example, hippocampal pyramidal neurons) express high levels of NMDA receptors3, that overactivation of glutamate receptors in cultured neurons and adult rats can elicit antigenic changes in the cytoskeleton similar to those seen in neurofibrillarly tangles in AD (ref. 21), and that Aβ increases neuronal vulnerability to excitotoxicity22,23. The data indicate two possible mechanisms whereby PS1 mutations might increase neuronal vulnerability to excitotoxicity. By increasing production of neurotoxic Aβ (refs. 11–13)and decreasing production of

Fig. 5 Agents that stabilize calcium homeostasis and antioxidants counteract the excitotoxicity-enhancing action of mutant PS1. Hippocampal cultures from wild-type mice (WT PS-1) and PS1M146VKI mice were pretreated for 2 h with either 1 µM dantrolene (DTL), 1 µM nifedipine (NIF), 200 µM uric acid (UA) or 5 µM propyl gallate (PG). Cultures were then exposed for 12 h to 50 µM glutamate, and neuron survival was quantified (Control cultures were not exposed to glutamate). Values are the mean and s.e.m. of determinations made in four cultures. **, P < 0.01; ***, P < 0.001 compared with value for cultures exposed to glutamate alone (ANOVA with Scheffe’s post-hoc tests). NATURE MEDICINE • VOLUME 5 • NUMBER 1 • JANUARY 1999


ARTICLES the neuroprotective secreted form of APP (refs. 24–26), PS1 mutations would disrupt calcium homeostasis. PS1 mutations also perturb calcium regulation in a manner that leads to increased release of calcium from the endoplasmic reticulum9,10. The endangering action of PS1 mutations is likely to sensitize the neurons to age-related changes in the brain including reduced energy availability and increased oxidative stress, conditions known to promote excitotoxic neuronal death1,2. Given evidence that PS1 mutations can increase the vulnerability of neural cells to apoptosis induced by trophic factor withdrawal and Aβ (refs. 9,10,26,27), our findings indicate that, by disturbing cellular calcium homeostasis, PS1 mutations predispose neurons to multiple forms of cell death. Preventive and therapeutic strategies that target both excitotoxic and apoptotic pathways may forestall the neurodegenerative process in AD.


Methods Generation of PS1 mutant knock-in mice. PCR primers (5’–GAGAGAGAAGGAA CCAACACAAGACAG–3’ and 5’–AGACATCTGGGCATTCTGGAAGTAGGA–3’) designed to amplify a 180-bp DNA fragment from exon 3 of the murine PS1 gene were used to isolate three P1 clones derived from a mouse 129/SvJ genomic library (Genome Systems, St. Louis, Missouri). An 8kb HindIII fragment (containing exons 4 and 5) and a 4-kb SacI fragment (containing exon 5) were subcloned from clone 292 (GS control number 8270), restriction mapped and partially sequenced. A 1.2-kb SacI–HindIII fragment (containing exon 5) was subcloned into pAlter-1 (Promega, Madison, Wisconsin) and mutagenized with a 39-bp mutagenic oligonucleotide (5’–ATGATCAGTGTCATTGTCGTGGTGACCATCCTCCTGGTG–3’) according to the manufacturer’s instructions. The mutagenic oligonucleotide introduced a I145V/M146V double mutation and a unique BstEII restriction enzyme site. 5’ and 3’ targeting arms were assembled in vectors derived from pZErO-2.1 (Invitrogen, Carlsbad, California) (containing newly engineered multiple cloning sites with a loxP site and appropriate restriction enzyme sites) and then subcloned into the targeting vector NTK2. The assembled targeting vector was linearized with PvuI and electroporated (at 240 V with a capacitance of 500 µF using a BioRad Gene Pulsar; BioRad, Richmond, California) into R1 embryonic stem cells that were derived from a 129/Sv × 129/Sv-CP F1 blastocyst. Genomic DNA from 250 clones surviving double selection with G418 (250 µg/ml) and ganciclovir (2 µM) was digested with HindIII and screened for homologous recombination by Southern blot analysis using a 5’ flanking probe (a 600-bp HindIII–HpaI fragment). Nineteen clones had the expected HindIII restriction fragment length polymorphism (caused by the insertion of the 5’ loxP site). All of the 19 targeted clones also showed the expected BglI restriction fragment length polymorphism (caused by the insertion of the neo slection cassette). Four of the targeted clones (clones 19, 106, 157 and 179) were microinjected into 3.5day-old blastocysts isolated from C57BL/6 mice, and were transferred to foster mothers to produce male chimeric mice. The contribution of these stem cells to the germline of chimeric mice was assessed by breeding with C57BL/6 females and screening for agouti offspring. Germline transmission of the PS1 mutation was achieved with all four of the injected embryonic stem cell lines. Mice heterozygous for the PS1 mutation were intercrossed to generate homozygous mutant knock-in mice (PS1M146VKI). Genotype assignments. PCR primers (5’–AGGCAGGAAGATCACGTGTTCAAGTAC–3’ and 5’–CACACGCACACTCTGACATGCACAGGC–3’) were used to amplify genomic DNA sequences flanking exon 5 before digestion of the amplified DNA with the restriction enzyme BstEII. After an initial ‘hot start’ at 94 °C for 2 min, 35 cycles (94 °C for 40 s, 62 °C for 40 s, 72 °C for 1 min) were run. The expected PCR product is 530 bp. Expected product sizes for the wild-type and targeted allele after BstEII enzyme digestion are 530 and 350/180 bp respectively. Because none of the injected embryonic stem cell lines showed evidence of secondary neo insertions by Southern blot analysis (data not shown), we have relied almost exclusively on this assay to genotype PS1M146VKI mice. Reverse transcription–PCR (RT–PCR). Brain tissue was homogenized in 3 ml of TRIzol (Life Technologies) and total brain RNA was extracted accordNATURE MEDICINE • VOLUME 5 • NUMBER 1 • JANUARY 1999

ing to manufacturer’s instructions. Total RNA (2 µg) was reverse transcribed and the cDNA subjected to PCR amplication using PCR primers (5’–TGCTCCAATGACAGAGATACCTGCACC–3’ and 5’–GATAAGAGCTGGAAAGAGAGTCTC–3’) designed to amplify the cDNA sequences encoding amino acids 1–287 of PS1 (sequences encoded by exons 3 through 8). After an initial ‘hot start’ at 94 °C for 2 min, 35 cycles (94 °C for 40 s, 55 °C for 40 s, 72 °C for 1 min) were done. The expected PCR product is 868 bp. Expected product sizes for the wild-type and targeted allele after BstEII enzyme digestion are 868 bp and 426/442 bp, respectively. Western blot analysis. Relative levels of PS1 expression were determined by western blot analysis using similar methods to those described9,10. Solubilized protein (50 µg) was separated SDS–PAGE and then transferred to a nitrocellulose sheet. After blocking with 5% milk and a 3-hour incubation in the presence of primary PS1 antibody9, the nitrocellulose sheet was further processed using horseradish peroxidase-conjugated secondary antibody and a chemiluminescence system (Amersham). Hippocampal cell cultures. Dissociated hippocampal cell cultures were prepared from postnatal-day-1 wild-type and homozygous PS1M146VKI mouse pups using methods similar to those described21. Hippocampi were removed and incubated for 15 min in Ca2+- and Mg2+-free Hank’s Balanced Saline Solution (Life Technologies) containing 0.2% papain. Cells were dissociated by trituration and plated into polyethyleneimine-coated plastic or glass-bottom culture dishes containing Minimum Essential Medium with Earle’s salts supplemented with 10% heat-inactivated fetal bovine serum, 2 mM L-glutamine, 1 mM pyruvate, 20 mM KCl, 10 mM sodium bicarbonate and 1 mM HEPES, pH 7.2. After cell attachment (3–6 h after plating), the culture medium was replaced with Neurobasal Medium with B27 supplements (Life Technologies). Experiments were done in 8-day-old cultures. Immediately before experimental treatment, the medium was replaced with Locke’s buffer (154 mM NaCl, 5.6 mM KCl, 2.3 mM CaCl 2, 1.0 mM MgCl2, 3.6 mM NaHCO3, 5 mM glucose, 5 mM Hepes, pH 7.2). Experimental treatments and quantification of neuronal survival. Glutamate was prepared as 200× stocks in Locke’s buffer. Nifedipine and dantrolene (Sigma), and BAPTA-AM (Molecular Probes, Eugene, Oregon) were prepared as 500× stocks in dimethylsulfoxide. The method for quantification of neuron survival in hippocampal cell cultures has been described20. Undamaged neurons in pre-marked microscope fields were counted before and at different time points after exposure to experimental treatments. A neuron with intact neurites and a cell body that was smooth and round to oval in shape was considered viable. A neuron with beaded or fragmented neurites and a cell body that was shrunken and rough in appearance was considered nonviable. Measurement of intracellular calcium levels. Cytoplasmic free calcium levels were quantified by fluorescence ratio imaging of the calcium indicator dye fura-2 using methods described9,10. Cells were loaded with the acetoxymethylester form of fura-2 (30 min incubation in the presence of 10 µM fura-2) and imaged using a Zeiss AttoFluor system with a 40× oil objective. The average intracellular calcium concentration ([Ca2+]i) in individual neuronal cell bodies was determined from the ratio of the fluorescence emissions obtained using two different excitation wavelengths (334 nm and 380nm). The system was calibrated using solutions containing either no calcium or a saturating level of calcium (1 mM) using the formula: [Ca 2+]i = Kd[(R–Rmin)/(Rmax–R)](Fo/Fs)(ref. 28). Assessments of oxidative stress and mitochondrial transmembrane potential. Levels of intracellular peroxide levels were measured using the probe 2,7-dichlorofluorescin diacetate, which is oxidized to fluorescent compound 2,7-dichlorofluorescein by peroxides, using methods similar to those described20. Cells were incubated for 30 min in the presence of 10 µM 2,7dichlorofluorescin diacetate (Molecular Probes, Eugene, Oregon), washed twice with Locke’s solution, and confocal images of cell-associated 2,7dichlorofluorescein fluorescence were acquired (488 nm excitation and 510 nm emission). The average pixel intensity in individual cell bodies was determined using Imagespace software (Molecular Dynamics, Sunnyvale, California); all images were coded and analyzed by experimenters without knowledge of experimental treatment history of the cultures. The dye dihy105


ARTICLES


drorhodamine, which localizes to mitochondria and fluoresces when oxidized to the positively charged rhodamine 123 derivative, was used to quantify relative levels of mitochondrial reactive oxygen species using methods similar to those described29. Cells were incubated for 30 min in the presence of 5 µM dihydrorhodamine, washed three times with Locke’s solution, and confocal images of cellular fluorescence were acquired and analyzed as described for 2,7-dichlorofluorescein fluorescence. The thiobarbituric reactive substances assay, which measures relative levels of membrane lipid peroxidation, was done as described30. The dye rhodamine 123 (Molecular Probes, Eugene, Oregon) was used as a measure of mitochondrial transmembrane potential using methods described10. Cells were incubated for 30 min in the presence of 10 µM of the dye, washed twice in Locke’s solution, and confocal images of cellular fluorescence were acquired and analyzed as described for ethidium fluorescence. Kainate administration and assessments of synaptic degeneration and neuronal death in vivo. These methods were similar to those used previously18. KA (0.3 µg in a volume of 0.5 µl) was injected unilaterally into dorsal hippocampus (from bregma: dorso-ventral, –2.0, medio-lateral, +2.4; anterio-posterior, –1.8) of anesthetized male mice 3 months old and 25–30 g in body weight. All mice given KA had seizures within the first hour after injection. Mice were killed 24 h later and perfused transcardially with 4% paraformaldehyde. Coronal brain sections (30 µm in thickness) were cut on a freezing microtome and used for Nissl staining and immunohistochemistry. Nissl-positive undamaged neurons were counted in hippocampal regions CA1, CA3 and CA4/hilus (counts were made in three 40× fields and four sections per brain). For electron microscopic examination, the tissues were perfused with 4% paraformaldehyde, 2% glutaraldehyde. The tissues were then washed in 0.1 M Sorenson’s phosphate buffer, post-fixed in 0.1% osmium tetroxide, dehydrated in graded series of ethanol (50% through absolute), cleared in propylene oxide, infiltrated overnight with 50/50 propylene oxide/resin, and embedded in Spurrs resin for 48 h at 60 °C. Samples were then sectioned on a Reichert-Jung ultramicrotome and examined in the Hitachi H7000 transmission electron microscope. Cell counts of Nissl-stained sections and electron microscope analyses were done by experimenters without knowledge of the genotype of the mice. Acknowledgments We thank J.N. Keller and A.J. Bruce-Keller for discussions. This work was supported by grants to M.P.M. from the NIH (NIA and NINDS), to C.B.W. from the NIA, to B.L.S. from the University of Washington Nathan Shock Center for Excellence in the Basic Biology of Aging, and to G.M.M. from the NIA.

RECEIVED 22 OCTOBER; ACCEPTED 19 NOVEMBER 1998 1. Olney, J. W. & DeGubareff, T. Glutamate and the pathophysiology of hypotoxicishemic brain damage. Ann. Neurol. 271, 557–559 (1978). 2. Choi, D.W. Calcium: still center stage in hypoxic-ischemic neuronal death. Trends Neurosci. 18, 58–60 (1996). 3. Greenamyre, J.T. & Young, A.B. Excitatory amino acids and Alzheimer’s disease. Neurobiol. Aging 10, 593–602 (1989). 4. Mattson, M.P. et al. β-amyloid precursor protein metabolites and loss of neuronal calcium homeostasis in Alzheimer’s disease. Trends Neurosci. 16, 409–415 (1993). 5. Hardy, J. Amyloid, the presenilins and Alzheimer’s disease. Trends Neurosci. 20, 154–159 (1997).

106

6. Kim, T.W. & Tanzi, R.E. Presenilins and Alzheimer’s disease. Curr. Opin. Neurobiol. 7, 683–688 (1997). 7. Zhou, Q., Quaife, C.J. & Palmiter, R.Î. Targeted disruption of the tyrosine hydroxylase gene reveals that catecholoamines are required for mouse fetal development. Nature 374, 640–643 (1995). 8. Nagy, A. et al. Derivation of completely cell culture-derived mice from early-passage embryonic stem cells. Proc. Natl. Acad. Sci. USA 90, 8424–8428 (1993). 9. Guo, Q. et al. Alzheimer’s presenilin mutation sesnsitizes neural cells to apoptosis induced by trophic factor withdrawal and amyloid β-peptide: involvement of calcium and oxyradicals. J. Neurosci. 17, 4212–4222 (1997). 10. Guo, Q., Christakos, S., Robinson, N. & Mattson, M.P. Calbindin blocks the proapoptotic actions of mutant presenilin-1: reduced oxidative stress and preserved mitochondrial function. Proc. Natl. Acad. Sci. USA 95, 3227–3232 (1998). 11. Scheuner, D. et al. The amyloid β protein deposited in the senile plaques of Alzheimer’s disease is increased in vivo by the presenilin 1 and 2 and APP mutations linked to familial Alzheimer’s disease. Nature Med. 2, 864–870 (1996). 12. Borchelt, D.R. et al. Familial Alzheimer’s disease-linked presenilin 1 variants elevate Aβ1-42/1-40 ratio in vitro and in vivo. Neuron 17, 1005–1013 (1996). 13. Duff, K. et al. Increased amyloid-β42(43) in brains of mice expressing mutant presenilin 1. Nature 383, 710–713 (1996). 14. Wong, P.C. et al. Presenilin 1 is required for Notch1 and Dll1 expression in the paraxial mesoderm. Nature 387, 288–292 (1997). 15. Shen, J. et al. Skeletal and CNS defects in Presenilin-1-deficient mice. Cell 89, 629–639 (1997). 16. Qian, S. et al. Mutant human presenilin 1 protects presenilin 1 null mouse against embryonic lethality and elevates Aβ1-42/43 expression. Neuron 20, 611–617 (1998). 17. Davis, J.A. et al. An Alzheimer’s disease-linked PS1 variant rescues the developmental abnormalities of PS1-deficient embryos. Neuron 20, 603–609 (1998). 18. Bruce, A. J. et al. Altered neuronal and microglial responses to brain injury in mice lacking TNF receptors. Nature Med. 2, 788–794 (1996). 19. White, R.J. & Reynolds, I.J. Mitochondrial depolarization in glutamate-stimulated neurons: an early signal specific to excitotoxin exposure. J. Neurosci. 16, 5688–5697 (1996). 20. Mattson, M.P., Lovell, M.A., Furukawa, K. & Markesbery, W.R. Neurotrophic factors attenuate glutamate-induced accumulation of peroxides, elevation of intracellular Ca2+ concentration, and neurotoxicity and increase antioxidant enzyme activities in hippocampal neurons. J. Neurochem. 65, 1740–1751 (1995). 21. Mattson, M.P. Antigenic changes similar to those seen in neurofibrillary tangles are elicited by glutamate and calcium influx in cultured hippocampal neurons. Neuron 4, 105–117 (1990). 22. Koh, J., Yang, L.L. & Cotman, C.W. β-amyloid protein increases the vulnerability of cultured cortical neurons to excitotoxic damage. Brain Res. 533, 315–320 (1990). 23. Mattson, M. P. et al. β-amyloid peptides destabilize calcium homeostasis and render human cortical neurons vulnerable to excitotoxicity. J. Neurosci. 12, 376–389 (1992). 24. Mattson, M.P. et al. Evidence for excitoprotective and intraneuronal calcium-regulating roles for secreted forms of β-amyloid precursor protein. Neuron 10, 243–254 (1993). 25. Ancolio, K., Marambaud, P., Dauch, P. & Checler, F. α-secretase-derived product of β-amyloid precursor protein is decreased by presenilin 1 mutations linked to familial Alzheimer’s disease. J. Neurochem. 69, 2494–2499 (1997). 26. Guo, Q., Robinson, N. & Mattson, M.P. Secreted APPα counteracts the pro-apoptotic action of mutant presenilin-1 by activation of NF-κB and stabilization of calcium homeostasis J. Biol. Chem. 273, 12341–12351 (1998). 27. Wolozin, B. et al. Participation of presenilin 2 in apoptosis: enhanced basal activity conferred by an Alzheimer mutation. Science 274, 1710–1713 (1996). 28. Grynkiewicz, G., Poenie, M. & Tsien, R.Y. A new generation of calcium indicators with greatly improved fluorescence properties. J. Biol. Chem. 260, 3440–3450 (1985). 29. Keller, J.N. et al. Increased sensitivity to mitochondrial toxin-induced apoptosis in neural cells expressing mutant presenilin-1 is linked to perturbed calcium homeostasis and enhanced oxyradical production. J. Neurosci. 18, 4439–4450 (1998). 30. Goodman, Y. & Mattson, M.P. Ceramide protects hippocampal neurons against excitotoxic and oxidative insults, and amyloid b-peptide toxicity. J. Neurochem. 66, 869–872 (1996).

NATURE MEDICINE • VOLUME 5 • NUMBER 1 • JANUARY 1999