| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Molecular Neuropsychiatry Branch, National Institute on Drug Abuse-Intramural Research Program, National Institutes of Health, DHHS, Baltimore, Maryland, USA
1 Correspondence: Molecular Neuropsychiatry Branch, National Institute on Drug Abuse, Intramural Research Program, National Institute of Health, DHHS, 5500 Nathan Shock Dr., Baltimore, MD 21224, USA. E-mail: jcadet{at}intra.nida.nih.gov
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: apoptosis-inducing factor • CHOP transcription factor • BiP/GRP78 • DNA fragmentation
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
2
3)
, it is essential to study the molecular and cellular mechanisms of methamphetamine (METH) toxicity. For example, acute METH intoxication can result in belligerent and psychotic behavior (4)
as well as multiple organ failure (5
, 6)
in humans. Myocardial infarction, stroke, cerebral hemorrhages, and death have also been reported (6)
. Dysfunctions in dopaminergic and nondopaminergic regions of the brains of METH abusers have been observed in various imaging studies (2
, 3
, 7
8
9)
. Some of the results of the imaging studies are supported by a postmortem study that showed marked decreases in the levels of striatal dopamine, dopamine transporters, and serotonin in the brains of METH-addicted patients (1)
.
In animals, METH causes substantial brain damage characterized by decreases in striatal dopamine and serotonin levels, decreased tyrosine hydroxylase activity, and loss of dopamine transporters (see reviews, refs 10
11
12
). Although studied less, incontrovertible evidence has long existed that METH can cause cell death in the brain (13
14
15
16)
. Zalis et al. (13)
injected dogs with 2.5–5 mg/kg i.v. or 10 mg/kg p.o and reported the occurrence of neuronal cell degeneration in the cerebral cortex, basal ganglia, brainstem, and cerebellum of these animals. Ellinwood and collaborators (14)
had reported chromatolysis in various regions of the brains of cats treated with METH (2 wk of chronic intoxication with increasing doses of METH starting at 15 mg/kg and reaching 28 mg·kg–1·day–1). Subsequently, Ellison and Switzer (16)
demonstrated that METH (6 mg/kg given every 2 h for 8 h) caused pronounced degeneration in the striatum and cerebellum of rats killed 36 h after drug administration. More recently, Schmued and Bowyer (17)
reported that neuronal cell death was apparent in the hippocampal remnants of rats within 5 days after injections of METH (15 mg/kg x 4 administered at 2 h intervals). Eisch et al. (18
, 19)
had shown that METH (4 mg/kg x 4 given at 2 h intervals) caused cell death that peaked at 
3 days after drug administration. In a series of recent papers, we have demonstrated that METH-induced cell death showed characteristics of apoptosis (20
21
22
23
24
25
26)
, findings that have been replicated by others (27
28
29)
.
Apoptosis is a highly regulated death process that occurs during development and is thought to be dysregulated in several neurodegenerative disorders including Parkinson’s disease, Huntington’s chorea, amyotrophic lateral sclerosis, and Alzheimer’s disease (30)
. Neuronal apoptosis can be induced by stimulation of plasma membrane death receptors and by perturbation of intracellular homeostasis via activation of specific organelle-mediated death cascades (31)
. For example, damage to mitochondria and the endoplasmic reticulum (ER) (32
, 33)
can cause release of cytochrome c (cyto c) and subsequent activation of caspases that are major mediators of apoptotic signals. These enzymes are broadly divided into two groups: initiator caspases whose main function is to activate downstream caspases and effector/executor caspases responsible for dismantling cellular proteins. Activation of effector caspases leads to the proteolysis of several target proteins, including poly (ADP-ribose) polymerase (PARP), lamins and DNA fragmentation factor 45 kDa subunit (DFF-45) (34)
. The importance of proteolytic cleavage to the ensuing morphological and molecular changes associated with apoptotic phenomena is being actively investigated. DFF-40/CAD, a subunit of the heterodimeric DFF, has been shown to mediate genomic DNA degradation during apoptosis (35)
whereas PARP cleavage might cause dysfunctions in DNA repair mechanisms (36)
. Cleavage of lamins may interfere with the integrity of the nuclear envelope (37)
. Apoptosis can occur via caspase-independent mechanisms after the mitochondrial release of the apoptosis-inducing factor (AIF) (38)
.
Our efforts to characterize mechanisms involved in METH-induced neuronal apoptosis have provided some novel clues on the cellular and molecular effects of this illicit neurotoxin (12)
. We have found that METH-induced cell death is associated with the activation of the SAPK/JNK (stress-activated protein kinase/c-Jun amino-terminal kinase) pathway (39)
. Because the majority of TUNEL-positive cells express phosphorylated c-Jun and c-jun knockout mice show partial protection against METH-mediated neuronal apoptosis (23)
, it appears that the SAPK pathway plays a role in causing METH-induced cell death. These findings are consistent with a wealth of information that implicates c-Jun as a proapoptotic agent (40)
. We and others have found that METH-induced neurodegeneration is dependent on mitochondrial mechanisms in vivo (26)
and in vitro (22
, 27)
. Specifically, administration of METH to mice causes increases in pro-death (BAX, BAD, and BID) but decreases in antiapoptotic Bcl-2-related proteins (Bcl-2 and Bcl-XL) (26)
. These proteins are known to be involved in either activating or inhibiting the mitochondria-dependent cell death pathway in the mammalian brain (41)
. We report here that, in addition to the mitochondrial death pathway, METH can exert its neurodegenerative effects by causing ER stress in the striatal neurons of mice treated with toxic doses of the drug and that the mitochondria and ER pathways are activated within the same cells before experiencing their ultimate demise.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, 26
, 39
, 42)
and by other investigators (43
44
45
46
47)
. Animals showed no evidence of seizures and all mice survived this dose regimen throughout the duration of the study (7 days at most), during which they were killed at various times after drug treatment. As suggested by some investigators (45
46
47)
, this approach might help to more specifically define the biochemical and molecular bases of METH-induced neurotoxicity. The use of large doses of METH constitutes an attempt to mimic the large doses taken by human METH abusers, which can amount to several grams taken in a day (4
, 48)
. Because there are significant species differences in METH elimination half-lives (49)
and metabolism (50)
, the often used binging patterns introduced in 1988 by Sonsalla et al. (51)
are only approximations of what is actually encountered in clinical situations. This issue has been discussed extensively in a recent review of METH neurotoxicity (12)
. Brain tissues were processed for use in immunohistochemistry, Western blot, and real-time RT-PCR as described below. All animal use procedures were according to the NIH Guide for the Care and Use of Laboratory Animals and were approved by the local NIDA Animal Care Committee.
Immunocytochemistry and TUNEL histochemistry
Procedures were performed according to previously reported methods (24
, 25)
. At designated times after METH administration, the animals were perfused transcardially with 4% paraformaldehyde under deep pentobarbital anesthesia. The brain tissues were subsequently removed and postfixed overnight in the same fixative. On the next day, 30 µm coronal sections were cut using a cryostat. Sections were mounted onto Superfrost/Plus microscopy slides (Fisher, PA, USA) and stored at -20°C.
METH-induced striatal apoptotic cells and their content were detected by using dual antigen staining of sections with TUNEL histochemistry and neuronal nuclei (NeuN) or GAD immunohistochemistry as described before (23)
. Slide-mounted sections from saline- or 3 day METH-treated animals were incubated with anti-NeuN or anti-GAD primary antibody (both from Chemicon International Inc., Temecula, CA, USA). They were subsequently incubated with biotinylated secondary antibody and Texas red-avidin-DCS (both from Vector Laboratories, Burlingame, CA, USA). After NeuN or GAD immunostaining, the sections were processed by TUNEL histochemical staining as reported earlier by this laboratory (23
, 25)
. The immunostained sections were rinsed in 0.5% Triton X-100 in 0.01M phosphate-buffered saline for 20 min at 80°C to increase permeabilization of the cells. To label damaged nuclei, 50 µL of the TUNEL reaction mixture (Roche Diagnostics Corporation, IN, USA) were added onto each sample in a humidified chamber, followed by a 60 min incubation at 37°C. Procedures for negative controls were carried out as described in the manufacturer’s manual and consisted of not adding the label solution (terminal deoxynucleotidyl transferase) to the TUNEL reaction mixture. No TUNEL-positive cells were observed in the negative controls.
To detect the activation of caspase-3 in striatal neurons, we used costaining with antibodies against NeuN and activated caspase-3. Briefly, slide-mounted sections from saline- and METH-treated animals (8 and 24 h) were incubated with monoclonal anti-NeuN (Chemicon) and polyclonal anti-cleaved caspase-3 (Cell Signaling, Beverly, MA, USA) primary antibodies. This was followed by incubation with Texas red-conjugated anti-mouse antibody and biotinylated anti-rabbit antibody, then FITC-avidin-DCS immersion to detect the polyclonal antibody (caspase-3).
Dual antigen immunostaining of CHOP/GADD153, a marker for ER stress, and cleaved caspase-3 antibodies was performed as described above. After staining, images were processed by using a Carl Zeiss Laser Scanning Confocal System with Axiovert 135 inverted microscopy. Excitation and emission wavelengths were selected according to the suggested index. To estimate the percentages of cell death of neuronal or GABAergic origin, striata of mice (n=10) were randomly viewed under 20x objective lens. Total neurons dying (green) = a, total GAD neurons that are dying (yellow) = b, and total number of GABA cells (red) = c were counted (see Fig. 1
B for representative pictures). The percentage of GABA cells that die was calculated using the equation, 100 x b/c.
|
Western blot analysis
Immunoblot analysis was carried out with the striatum dissected from six mice of METH- and saline-treated wild-type mice. Samples from six mice were pooled to form one group. Experiments were performed from three different groups for quantitation. The pooled mouse striatum was homogenized in buffer containing 50 mM Tris-HCl (pH 7.5), 150 mM NaCl, 1 mM EGTA, 1 mM PMSF, 0.5% NP-40, 0.25% SDS, 5 µg/mL leupeptin, and 5 µg/mL aprotinin. Homogenates were centrifuged at 750 g for 10 min at 4°C and the precipitates that belong to nuclear fractions were resuspended in the lysis buffer. The supernatant fraction was subsequently centrifuged at 10,000 g for 15 min at 4°C. The resulting pellet belongs to mitochondrial fraction and was resuspended in the lysis buffer. Supernatants were further centrifuged at 100,000 g for 1 h at 4°C. The pellet was discarded and the remaining supernatant is the cytosolic fraction. After determining the protein concentration of lysates with a Bio-Rad assay system (Bio-Rad, San Francisco, CA, USA), the lysates were denatured with sample buffer (62.5 mM Tris-HCl, 10% glycerol, 2% SDS, 0.1% bromophenol blue, and 50 mM DTT) at 100°C for 5 min and subjected to SDS-PAGE. Proteins were electrophoretically transferred to Hybond-PTM membrane (Amersham Pharmacia Biotech. Piscataway, NJ, USA). Membrane blocking, primary and secondary antibody incubations, and chemiluminescence reactions were carried out according to the protocol described by individual antibody suppliers. Antibodies included monoclonal for Apaf-1, cyto c (BD PharMingen, San Diego, CA, USA), rabbit polyclonal caspase 3 for Western blot, DFF-40 and DFF-45 (BD PharMingen), rabbit polyclonal for Smac/DIABLO, caspases 9, 6, cleaved caspase 3 for immunohistochemistry, cleaved PARP and cleaved lamin A (Cell Signaling), rabbit polyclonal GAD-67, mouse monoclonal CHOP/GADD153 (Santa Cruz Biotechnology, Inc., Santa Cruz, CA, USA), rabbit polyclonal calpain, caspase 12 and AIF (Biovision Research Products, Mountain View, CA, USA), and rabbit polyclonal BiP/GRP78 (Stressgen Biotechnologies, Victoria, Canada). To confirm equal protein loading, blots were reprobed with 
-tubulin antibody (1:2000; Sigma, 2 h at RT). Signal intensity was measured using densitometric analysis (IS-1000 Digital Imaging System, Alpha InnoTech Corp., San Leandro, CA, USA) and quantitated using FluorChem version 2.0 software (AlphaEaseFC analysis software) (26)
.
Reverse transcription (RT) -PCR and detection of mRNA expression using LightCycler technique
Real-time PCR to detect mRNA expression for GAD67 and the ER stress-related genes chop and Grp78/BiP was done using the LightCycler thermal cycler system (Roche Diagnostics) essentially as described by us (52)
. For real-time PCR, unpooled total RNA (1 µg) from 6 to 10 CD1 mice per group was reverse-transcribed with oligo dT primer using Advantage RT for PCR kit (BD Biosciences Clontech Laboratories, Palo Alto, CA, USA). PCR experiments were then performed using light cycler technology and the LightCycler FastStart DNA Master SYBR Green I kit (Roche Molecular Biochemicals). HPLC-purified and gene-specific primers corresponding to PCR targets were obtained from the Synthesis and Sequencing Facility of Johns Hopkins University (Baltimore, MD, USA) (Table 1
).
|
For PCR, 2 µL of template was placed into 18 µL reaction volume containing 0.5 µL of each primer, 2.4 µL 25 mM MgCl2, 2 µL FastStart DNA Master SYBR Green I, and 13.6 µL dH2O. Nucleotides, Taq DNA polymerase, and buffer were included in the FastStart DNA Master SYBR Green I mix (Roche Diagnostics). A typical protocol took 60 min to complete and included denaturation at 95°C with a preincubation time of 8 min, followed by 45 cycles with 95°C denaturation for 15 s, 64°C annealing for 5 s, and 72°C extension for 15 s. Extension periods varied with the specific primers, depending on the length of the product. Negative controls were run concomitantly to confirm the overall specificity and to verify that no primer/dimer was generated. The relative standard curve was established with serial dilution of a cDNA solution with an unknown concentration that corresponds to a mix of five randomly picked samples. To confirm the amplification specificity, the PCR products were subjected to a melting curve analysis. Amplification curves were generated by the LightCycler Instrument Quantification program and displayed the fluorescence values vs. cycle number. Template concentrations using the relative standard curve were given arbitrary values. The mean concentration of clathrin light chain was used as control for input RNA because it is considered a stable housekeeping gene. The mean clathrin concentration was determined once for each cDNA sample and used to normalize all other genes tested from the same cDNA sample. The relative change in gene expression was recorded as the ratio of normalized data over saline. One-way ANOVA was followed by Fisher’s PLSD test for testing differences between the different time points and the saline-treated animals. All analyses were done using the program Statview 4.02 (SAS Institute, Cary, NC, USA). The null hypothesis was rejected at P < 0.05.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
shows the effects of METH on NeuN-positive striatal cells after 3 days of METH administration. The majority of the TUNEL-positive cells were NeuN positive (Fig. 1A
). The striatum is a complex structure of several subtypes of neurons, 95% of which are GABAergic neurons that project to the outside of that structure (53)
. We thus wanted to ascertain whether the affected neurons were GABAergic by running double-label experiments using an antibody against glutamic acid decarboxylase (GAD), the enzyme involved in the GABA synthetic pathway (54)
, in conjunction with TUNEL histochemistry. Figure 1B
shows that many of the dying cells were GAD positive, indicating that METH causes death of striatal GABAergic neurons. Cell counting established that 17% of the GAD-positive cells were TUNEL-positive. To test the effects of METH on GAD expression, we performed quantitative RT-PCR and Western blot analysis using primers and antibody specific for GAD67. Figure 1C
shows that METH caused a marked reduction in GAD67 transcript and protein levels measured several days after administration of the drug.
METH induces release of apoptogenic molecules from mitochondria
Because release of cyto c from mitochondria into the cytoplasm has been documented in apoptosis caused by several toxic triggers (55)
, we wanted to know whether METH injections to mice would cause similar changes in the compartmentalization of cyto c. To test this, we performed a detailed time course assessment of the concentration of cyto c in cellular subfractions. There was only a negligible amount of cyto c in the cytosolic fractions obtained from control mice. Administration of METH caused a gradual appearance of cyto c in the cytosol of striatal cells, which began at 30 min to 1 h, peaked at 4–8 h, then tapered off 2 days after drug injection (Fig. 2
A). These changes were contrasted by marked decreases in cyto c in mitochondrial fractions, with its almost total disappearance from the mitochondria from 2 to 7 days postdrug. Because cyto c is known to bind to dATP and apoptotic protease-activating factor-1 (Apaf-1) in order to form the apoptosome that activates caspase-9 (56)
, we determined whether METH would affect Apaf-1 and caspase-9 protein levels in the striatum. As is observed in Fig. 3
A, A’, Apaf-1 showed initial increases at 4 h after METH administration, with maximal changes occurring during the 8–24 h interval after the drug injection. These were accompanied by initial caspase-9 cleavage at 4 h, with the cleaved protein reaching a peak concentration between 8–24 h postdrug (Fig. 3A, A’
).
|
|
Besides cyto c, several other apoptogenic substances can be released from the mitochondria during the apoptotic process. These include Smac/DIABLO (57)
and AIF (58)
. We measured these two substances in order to test the possibility that METH administration can release proteins that might be involved in caspase-independent effects in the brain. Figure 2B, B’
shows that METH induced transit of Smac/DIABLO from the mitochondria to the cytosol as early as 30 min to 1 h after treatment. In addition, METH caused AIF, which has been reported to cause caspase-independent apoptosis (59)
, to be released from the mitochondria (30 min–1 h) and to be translocated to the nucleus (8–72 h) via transit through the cytosol (Fig. 2C, C’
). These findings are consistent with a recent paper in which similar AIF transit was observed during apoptosis caused by N-methyl-D-aspartate (NMDA) glutamate receptor-mediated toxicity in cortical neurons (60)
.
Neuronal apoptosis is accompanied by cleavage of caspases-3 and -6 but not of caspase-7
Because activation of the mitochondrial death pathway is known to lead to cleavage of terminal caspases involved in the proteolysis of many proteins (34)
, we sought to determine whether this was the case with METH treatment. There was indeed cleavage of the caspase-3 pro-enzyme (32 kDa), with the cleaved (17 kDa) product appearing and increasing from 3 to 48 h after drug administration (Fig. 3B, B’
). To determine whether caspase-3 cleavage was occurring in striatal neurons, we performed double-label immunostaining experiments using antibodies against active caspase-3 and NeuN. As shown in Fig. 4
, METH-induced changes in the activation of caspase-3 occurred in NeuN-positive cells in the mouse striatum. Figure 3C
shows that METH administration caused cleavage of pro-caspase-6, beginning at 16 h and almost completely disappearing at 7 days post-METH treatment. In contrast, METH caused no changes in the expression of caspase-7 (data not shown).
|
METH administration causes ER stress in the mouse striatum
The ER is an important organelle that participates in cellular homeostasis by regulating calcium signaling and protein folding (61)
. However, dysregulation of intracellular homeostasis (62)
and oxidative stress (63)
can cause ER stress and ER-induced apoptosis (62)
. Because METH is known to cause oxidative stress (10
, 64)
, we wanted to know whether ER-mediated events might participate in METH-induced cellular demise. We measured a number of substances that have been reported to participate in ER-induced apoptosis and the unfolded protein response (UPR) (62)
. These include caspase-12, calpain (65)
, GRP78/BiP (glucose-regulated protein/immunoglobulin heavy chain binding protein) (66)
, and CHOP/GADD153 (C/EBP homology protein/growth arrest and DNA damage) (67)
. As shown in Fig. 5
A, METH administration caused marked changes in the pattern of caspase-12 expression. Specifically, 1 h after METH injection there was an initial cytosolic appearance of the smaller cleaved 40 kDa fragment that peaked at 3–4 h, then gradually decreased toward control levels 7 days after drug treatment (Fig. 5A’
). Because caspase-12 is cleaved by calpain during ER stress (65)
, we assessed the effects of METH on calpain expression by using an antibody that detects the original and activated forms of this protease. We found that METH administration was associated with activation of calpain as early as 30 min after drug injection (Fig. 5B, B’
). These changes lasted throughout the time course of the experiments (7 days), suggesting the possible occurrence of METH-induced prolonged changes in calcium homeostasis since calpain is activated by increased intracellular calcium (68)
.
|
The ER stress-generated UPR is characterized by increased transcription of ER chaperones, such as BiP/GRP78, which are meant to increase the ER capacity to cope (66)
. To confirm that METH was indeed causing ER stress, we decided to measure BiP mRNA and protein expression after drug treatment. Figure 6
A (upper panel) shows that METH indeed caused very early increases (2- to 2.5-fold) in BiP/Grp78 mRNA at 30 min–2 h after its injection. Western blot analysis revealed a gradual increase in BiP protein starting at 8 h and reaching a maximal peak by 2–3 days after the METH injection (Fig. 6A
, lower panel).
|
In addition to increasing its coping mechanisms, organisms can trigger apoptosis by increasing the transcription of genes that are involved in causing cellular demise if the ER stress is overwhelming (reviewed in ref 69
). One such gene is chop/gadd153 (67)
, which inhibits some (63)
but induces other genes (70)
. We found that METH-treated mice showed rapid increases in chop mRNA that lasted from 30 min to 2 h (Fig. 6B
, upper panel). Figure 6B
(lower panel) shows that METH treatment caused sustained increases in CHOP protein from 4 h for up to 7 days after drug treatment. To examine whether ER stress was occurring in the same neurons that showed expression of activated caspase-3, we performed double-label experiments with antibodies against CHOP and active caspase-3. Figure 6C
shows that these two proteins colocalize in all affected cells.
METH administration causes cleavage of various caspase target proteins
Because activation of terminal caspases leads to the cleavage of several nuclear and cytoplasmic proteins, we investigated the effects of METH on the status of DFF45, DFF40, PARP, and lamin A, known caspase targets (34)
. Figure 7
A shows the effects of METH on the status of DFF45. In samples from control mice, there were two protein bands: one of 45 kDa and one of 34 kDa. METH injection caused decreases in the cytosolic DFF45 kDa band but increases in the 34 kDa band. There was also the appearance of METH-induced 32 and 11 kDa bands first observed at 8 h after drug treatment and present almost concurrently (Fig. 7A
). The concentration of the carboxyl-terminal 11 kDa fragment (p11), essential for activation of the DFF activity (71)
, peaked after 2 days and returned toward control levels by 7 days (Fig. 7A’
). Almost simultaneously, DFF40, the active component of DFF that triggers DNA fragmentation (72)
, showed increases in the cytosolic fractions after drug administration (Fig. 7A, A’
). Another caspase-3 target, PARP, was cleaved to its 89 kDa product in nuclear fractions (Fig. 7B
). Because caspase-dependent degradation of lamins is thought to facilitate nuclear dissolution during nuclear apoptosis (37)
, we measured its status after METH treatment. We found that METH administration resulted in a time-dependent cleavage of lamin A in nuclear fractions (Fig. 7C
). The 46 and 28 kDa lamin A fragments, which are indicative of proteolytic digestion, were quite visible 8 h after METH administration. These changes in lamin A peaked at 24 h after METH, then tapered off until 7 days after drug treatment (Fig. 7C’
). The time (24 h after drug) of greatest changes in lamin A status corresponds to the time of peaked caspase-6 cleavage (see Fig. 3C
). Unexpectedly, there were no changes in the pattern of lamin B expression in spite of our running these experiments with three different antibodies (data not shown).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
although, as discussed earlier, there was substantial evidence that METH can cause cell death in the brains of various animal species (see review by Davidson et al., ref 73
). Results from several laboratories have now confirmed that METH can indeed cause neuronal death in vitro (20
, 22
, 27
28
29)
and in vivo (17
, 18
, 24
25
26)
. This is the first demonstration, however, that METH can kill GAD-positive striatal neurons in vivo.
GAD is the rate-limiting enzyme in GABA biosynthesis (54)
and has been used to characterize GABAergic neurons in various regions of the brain (74)
. GAD exists as two major isoforms, GAD65 and GAD67, which are products of two different genes (75)
. GAD65 is membrane-associated whereas GAD67 is cytoplasmic (75)
; GAD65 is thought to be involved in short-term regulation of GAD activity whereas GAD67 is thought to participate in its long-term regulation (76)
. GAD67 is responsible for the majority of GABA synthesis within the brain because lack of GAD65, as observed in GAD65 knockout mice, does not change brain GABA content (77)
whereas significant decreases in GAD activity and GABA content are observed in the brains of GAD67 knockout mice (78)
. In the present study, we thus focus on studies of GAD67 and have found significant METH-induced decreases in GAD67 mRNA and protein levels. These results are consistent with the double label experiments where we found a substantial number of TUNEL-positive cells that are GAD-positive cells. Taken together, these observations suggest that many of the dying/dead cells are indeed GABAergic. Although these observations are not consistent with those of others who had reported that toxic doses of METH did not affect GABA systems in the brain (79
, 80)
, there are some differences between these studies and ours. Hotchkiss et al. (79)
injected 15 mg/kg x 4 of METH given 6 h apart for 24 h and measured GAD activity in rats killed from 6 h to 30 days after the last injection. In another study (80)
, GABA uptake was measured only 1 h after the 1 day binge of four injections of 10 mg/kg. In contrast, our observations revealed that decreases in GAD67 did not occur until 3 to 7 days after injection of a single large toxic dose of the drug (see Fig. 1
); these observations are consistent with our previous demonstration that the number of METH-induced TUNEL-positive cells peaked 3 days postdrug (24
, 25)
. This peak is followed by a decrease in the number of TUNEL-positive cells, suggesting that these cells had been removed by scavenger cells (81)
. Thus, the decreases in GAD67 protein levels 3–7 days after METH appear to correspond to the death of striatal GABAergic neurons caused by METH. It is thus likely that the discrepancies between our results and those of others (79
, 80)
are due to the doses of METH, the schedule of its administration, the time of animal death, and the different assays used in the studies. In what follows, we discuss the possible scenarios that might lead to the demise of these cells.
The mechanisms by which these GABAergic cells die subsequent to METH might include excitotoxic damage (for a review, see ref 82
) since toxic doses of METH are known to cause glutamate release in the striatum (83)
, production of free radicals (10
, 64
, 84
, 85)
; or perturbations of DNA repair mechanisms (86)
, and the activation of mitochondrial and ER death pathways (see below). The cytotoxic effects of glutamate are known to occur through receptor-mediated and receptor-independent events (87)
. Receptor-mediated mechanisms could involve glutamate-induced NO synthesis with subsequent production of reactive oxygen and nitrogen species that could lead to a shift in mitochondrial membrane potential with subsequent production of metabolic stress and the occurrence of cell death (82)
. Another major player in glutamate-mediated excitotoxicity leading to neuronal apoptosis involves NMDAR-induced excessive Ca2+ influx, which activates Ca2+/calmodulin-regulated protein phosphatase and causes the release of cytochrome c from mitochondria, subsequent activation of the caspase cascade, and cytoskeletal breakdown (88)
. The idea for the possible involvement of glutamate in METH-induced GABAergic cell death is supported by the presence of glutamate receptors on these neurons (89
, 90)
. Nevertheless, studies using specific glutamate receptor blockers need to be conducted to test this idea further.
Our repeated attempts to further characterize the molecular and cellular bases of METH-induced apoptosis have now revealed that ER and mitochondria-associated events are involved in causing the demise of these GABAergic neurons in a fashion similar to what occurs in other models of cell death (31
, 91)
. Specifically, we have shown that METH injections can promote a shift in the balance of pro-death/anti-death proteins of the BAX/Bcl-2-related family (26)
. These changes are known to trigger programmed cell death in various models of apoptosis (92)
. These earlier observations had led us to conclude that mitochondria-mediated caspase-dependent events might play important roles in METH-induced deleterious effects (22
, 26)
. Nevertheless, several other apoptogenic substances can be released from the mitochondria during the apoptotic process, one of which is AIF, which promotes cellular suicide in a caspase-independent manner (58)
. To test the role of AIF, we measured its exit from the mitochondria after METH and have provided the first detailed evidence for the concurrent involvement of multiple mitochondrial pathways that appear to interact to cause METH-induced neuronal death. Several lines of evidence had in fact suggested a role for mitochondrial dysfunction in METH-induced neurotoxicity. Gluck et al. (84)
had shown that METH can disrupt the electron transport chain by inhibiting complex I activity, an event associated with decreased ATP production. Burrows et al. (93)
had reported rapid METH-mediated decreases in cytochrome oxidase (complex IV) activity and striatal depletion of ATP stores in the rat brain, observations that are consistent with our recent observations of complex IV inhibition by METH using an in vitro system consisting of immortalized striatal cells (22)
. Further evidence that mitochondria may be involved in METH toxicity was provided by our previous report that BAX, a known proapoptotic agent that acts via cyto c release (33)
, is induced according to a time point that suggested a similar role of BAX in METH-induced cyto c release (22
, 26)
. These observations are consistent with data from other models of neuronal apoptosis in which mitochondrial mechanisms are involved. These include neuronal apoptosis associated with NO toxicity (94)
or exposure of cells to H2O2 (95)
, both of which have been shown to be involved in METH-induced toxicity (10
, 12
, 87)
. Cyto c release is followed by the formation of the apoptosome and subsequent cleavage of caspase-9 and caspase-3 (56)
. Nevertheless, in our model of METH-induced cell death, the fact that caspase-3 cleavage anteceded caspase-9 activation suggests that its initial cleavage might be secondary to mechanisms other than the formation of the apoptosome/caspase-9 complex. Such possibilities include ER-dependent events associated with the METH-induced activation of caspase-12.
As stated, damage to mitochondria is known to cause the release of another proapoptotic factor, AIF, which can activate caspase-independent neuronal apoptosis (38)
. Normally localized to the intermembrane compartment of mitochondria (96)
, AIF is released into the cytoplasm after induction of permeability transition pores (96)
. The early transit of AIF out of the mitochondria after METH may be due to a similar mechanism. In the present study, the release of AIF from the mitochondria was detected 30 min after METH and in the nuclear fraction 4 h thereafter. Rapid AIF release and its translocation to the nucleus, which preceded cyto c release from mitochondria, have been reported by others (58
, 59
, 97)
. In contrast, Cregan et al. (38)
reported that the release of AIF occurs later than cyto c in a model of p53-dependent cell death. In the present model, the peak of AIF release (1–2 h post-METH) appears to occur before peaked cyto c release (4–8 h postdrug). Thus, our data are more consistent with those of the former investigators (58
, 59
, 97)
. The reasons for these discrepant reports regarding the timing of AIF release are not clear, but suggest interesting questions that need to be addressed when trying to identify and explain the mechanisms involved in the efflux of proapoptotic substances from the mitochondria. In any case, after its transit into the nucleus, AIF might function as an endonuclease that causes DNA fragmentation (59
, 98)
, even though the possibility that AIF might promote the activity of endonucleases (such as endonuclease G) needs to be considered since these two appear to work in concert to cause DNA fragmentation in C. elegans (99)
. Because AIF entry into the nucleus precedes DFF45 cleavage in the METH apoptosis model, it is not farfetched to suggest that the earliest appearance of DNA damage after METH injection might occur in a caspase-independent fashion, with the destruction of the nucleus being dependent on the involvement on caspase-dependent and -independent pathways.
In addition to cyto c and AIF, METH injection caused the release of smac/DIABLO from the mitochondria, with the appearance of small amounts of the protein as early as 30 min postdrug injection. Smac/DIABLO is a proapoptotic molecule recently identified by two different groups (57
, 100)
. Smac/DIABLO functions as an indirect activator of caspases by inhibiting the inhibitor of apoptosis proteins (IAPs) (57
, 100)
. It has been suggested that the release of smac/DIABLO from the mitochondria during apoptosis is dependent on active caspases and occurs downstream of cyto c release (101)
. In contrast, we found early and potent METH-induced mitochondrial release of smac/DIABLO, which precedes caspases-3 and 9 activation. Moreover, Smac/DIABLO release was more coincident with AIF release in the METH model. These observations might implicate similar mechanisms for their release, mechanisms that may differ from cyto c release. Furthermore, the temporal sequence of smac/DIABLO release is consistent with its known action on the IAPs and supports its involvement in promoting the activation of various caspases in the present neurodegeneration model.
This is the first evidence that injections of METH can trigger ER stress in the mouse brain. The ER is an organelle actively involved in the synthesis and proper folding of proteins as well as in their transport via the Golgi apparatus to their ultimate destinations (91)
. Dysfunctions of calcium homeostasis, protein misfolding, or oxidative stress can cause ER stress and cell death (102)
. ER-induced apoptosis is associated with early calpain-dependent activation of caspase-12 (65)
, which can lead to caspase-3 cleavage. Calpain, a Ca2+-responsive cytosolic cysteine protease (65)
that is an important early mediator of ER-dependent cell death, is activated by METH, suggesting that METH might cause dysregulation of calcium mechanisms. This could have occurred secondary to METH-mediated oxidative stress (10
, 64
, 103)
. When taken together, these observations suggest that calpain might be acting to cleave caspase-12 in the METH neurodegeneration model in a fashion similar to that observed in other models of cell death (65)
. ER stress is associated with increased transcription of the ER resident chaperone GRP78/BiP (104)
and the nuclear protein CHOP (102)
. To assess whether METH-induced changes in these two proteins were occurring at transcriptional and translational levels in the METH toxicity model, we measured their mRNA and protein levels. We observed almost immediate increases in BiP mRNA levels. These changes might be related to METH-induced oxidative stress (10
, 64
, 103)
because BiP mRNA is induced in cultures of neurons exposed to oxidative insults (104)
. The increases in BiP protein levels might serve a protective function because BiP overexpression protects cells against apoptotic insults (105)
. In contrast, chop induction might play a role in promoting METH-induced apoptosis, as it has been reported that the brains of chop/gadd153 null mice exhibit a fourfold reduction in apoptosis that resulted from ER stress (67)
. This view is supported by our observation that cells that demonstrated METH-induced increased CHOP protein showed active caspase-3 expression. The link between CHOP and apoptosis is thought to occur via down-regulation of Bcl-2 expression and exaggerated production of reactive oxygen species (63)
. This supposition might provide a partial explanation for the METH-induced down-regulation of Bcl-2 (26)
and increased lipid peroxidation (64)
we observed in mice striata.
Activation of downstream or executioner caspases, including caspase-3, -6, -7, is the reported penultimate step in the induction of apoptosis. Their participation in neuronal cell death is thought to be influenced by cell types and specific apoptotic signals (106)
. The present observation of METH-induced activation of caspase-6, but not of caspase-7, indicates there might be selective activation of specific downstream caspases during METH-mediated neuronal apoptosis. These findings are consistent with the report of increased caspase-6 activity but not of caspase-7 in human neurons undergoing apoptosis after serum deprivation (106)
. Caspase-3 is the caspase most frequently reported to participate in various models of neuronal apoptosis, including that caused by oxidative stress (96
, 107)
. Thus, our present data extend a role for this enzyme to another model of neuronal apoptosis in vivo. Our results show that a panoply of proteins known to be targeted for proteolytic cleavage by these enzymes is affected after METH treatment. We find proteolysis of the caspase-6 substrate lamin A, whose cleavage has been reported to be necessary for complete condensation of DNA during apoptosis (108)
. The caspase-3 substrates ICAD/DFF 45 and PARP (34)
are cleaved during METH-induced apoptosis. DFF is comprised of DFF45/ICAD and DFF40/CAD subunits. Its cleavage by caspase-3 results in the liberation of the active DFF40, the major nuclease implicated in caspase-dependent DNA fragmentation (72)
. As stated earlier, our findings of a substantial increase in DFF-40 and in AIF suggest for the first time that METH-induced DNA fragmentation in the mouse brain (25)
might be the result of concerted efforts of various nucleases.
In summary, our observations indicate that the administration of METH to mice causes activation of several apoptotic pathways that have documented roles in neuronal apoptosis (31)
. Our results indicate that METH activates these death cascades in a sequential manner. We have summarized our observations in a theoretical schema that seeks to detail the temporal sequence of these molecular events (Fig. 8
). Early activation of calpain and caspase-12 indicates that METH-induced ER stress might be the earliest contributor to the appearance of apoptosis in the mouse brain after administration of the drug. This is followed by activation of mitochondria-mediated events that trigger both caspase-dependent and independent pathways. Our observations are beginning to shed more light on the diverse pathways that are involved in METH-induced neuronal apoptosis and to identify ER- and mitochondria-mediated events as important causative culprits in METH-induced neurodegeneration in the mouse brain. Finally, further investigations are needed to characterize the role, if any, of dopamine, glutamate, and/or temperature regulation in METH-induced apoptosis in the striatum because these have all been implicated in METH-mediated toxic effects on monoaminergic systems (12
, 73
, 109)
.
|
Received for publication March 26, 2003. Accepted for publication October 22, 2003.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
