Address correspondence and reprint requests to Shin Kwak, Department of Neurology, Graduate School of Medicine, University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113–8655, Japan. E-mail: firstname.lastname@example.org
AMPA receptor-mediated neurotoxicity is currently the most plausible hypothesis for the etiology of amyotrophic lateral sclerosis (ALS). The mechanism initiating this type of neuronal death is believed to be exaggerated Ca2+-influx through AMPA receptors, which is critically determined by the presence or absence of the glutamate receptor subunit 2 (GluR2) in the assembly. We have provided the first quantitative measurements of the expression profile of AMPA receptor subunits mRNAs in human single neurons by means of quantitative RT–PCR with a laser microdissector. Among the AMPA subunits, GluR2 shared the vast majority throughout the neuronal subsets and tissues examined. Furthermore, both the expression level and the proportion of GluR2 mRNA in motoneurons were the lowest among all neuronal subsets examined, whereas those in motoneurons of ALS did not differ from the control group, implying that selective reduction of the GluR2 subunit cannot be a mechanism of AMPA receptor-mediated neurotoxicity in ALS. However, the low relative abundance of GluR2 might provide spinal motoneurons with conditions that are easily affected by changes of AMPA receptor properties including deficient GluR2 mRNA editing in ALS.
glutamate receptor subunit 2 unedited at the Q/R site
glutamate receptor subunit 2 edited at the Q/R site
cerebellar granule cells
cerebellar molecular layer
cortical pyramidal neurons
spinal dorsal horn neurons in substantia gelatinosa
spinal ventral gray matter
spinal ventral funiculus
cerebral white matter
Amyotrophic lateral sclerosis (ALS) is a progressive neurodegenerative disease characterized by a selective loss of upper and lower motoneurons and muscle wasting with pyramidal tract signs. Approximately 5–10% of all ALS cases are familial (FALS) with two causal genes that have been identified. One consists of point mutations in the copper/zinc superoxide dismutase 1 (Cu/Zn SOD1) gene (Rosen et al. 1993) and the other has deletion mutations in the ALS2 gene (Hadano et al. 2001; Yang et al. 2001). On the other hand, the etiology of sporadic ALS, which accounts for the majority of all ALS cases, remains to be resolved.
One of the most plausible hypotheses for selective neuronal death in sporadic ALS is excitotoxicity mediated by α-amino-3-hydroxy-5-methyl-4-isoxazolepropionate (AMPA) receptors, which are a subtype of ionotropic glutamate receptors. In fact, it has been repeatedly shown that motoneurons are differentially more vulnerable to AMPA receptor-mediated slow neuronal death than other neuronal subsets in rat and mouse spinal cord cultures (Rothstein et al. 1993; Vandenberghe et al. 2000a,b) and in the adult rat spinal cord (Nakamura et al. 1994; Kwak and Nakamura 1995). Mechanisms underlying this slow neuronal death have not been elucidated, but an exaggerated Ca2+ influx through activated AMPA receptor-coupled channels appears to play a key role (Carriedo et al. 1996; Lu et al. 1996). Indeed, the role of Ca2+-permeable AMPA receptors on neuronal cell death has been demonstrated in cultured hippocampal neurons of glutamate receptor subunit 2 (GluR2)-null mice (Iihara et al. 2001).
Functional AMPA receptors are homo- or hetero-oligomeric assembly that is composed, in various combinations, of four subunits among GluR1, GluR2, GluR3, and GluR4. The Ca2+ conductance of AMPA receptors markedly differs depending on whether GluR2 is among the composite. AMPA receptors that contain at least one GluR2 in their composite have low Ca2+ conductance, whereas those lacking GluR2 are Ca2+-permeable (Hollmann et al. 1991; Verdoorn et al. 1991; Burnashev et al. 1992; Koh et al. 1995). These properties of GluR2 are generated post-transcriptionally by RNA editing at the Q/R site in the putative second transmembrane domain, whereby the glutamine (Q) codon is substituted by the arginine (R) codon (Sommer et al. 1991).
Accordingly, two mechanisms have been proposed for the selective death of spinal motoneurons in ALS. One is the selective reduction of GluR2 (GluR2 hypothesis; Pellegrini-Giampietro et al. 1997) that results in a low relative abundance of GluR2 among the four subunits. The other is the reduction in GluR2 RNA editing that results in a relative increase of GluR2(Q) compared to GluR2(R). As to the latter hypothesis, we have reported that GluR2 RNA editing was specifically reduced in the ventral gray matter of ALS spinal cords (Takuma et al. 1999). However, only indirect evidence has been available about the former, which includes in situ hybridization or immunocytochemistry studies that reported the absence of GluR2 in spinal motoneurons of control subjects (Williams et al. 1997; Bar-Peled et al. 1999; Meyer et al. 1999) without evidence in ALS patients.
AMPA receptor subunit composition was not different between motoneurons and dorsal horn neurons in the rat dissociated spinal cord culture (Vandenberghe et al. 2000b). Consequently, it is necessary to investigate whether the selective neuronal death of ALS spinal motoneurons is the natural extension of differential motoneuronal vulnerability of the animal spinal cord.
Therefore, to explore the relevance of the change in Ca2+ conductance of AMPA receptors on ALS motoneurons, by means of a laser microdissector and real-time quantitative RT–PCR, we compared the following in this study: (i) the relative abundance of GluR2 mRNA in situ in human spinal motoneurons with that in other neuronal tissues from control subjects and (ii) the expression level of GluR2 mRNA in single motoneurons between ALS and control cases.
Materials and methods
Spinal cord and brain samples
Spinal cord and brain tissues were rapidly frozen on dry ice immediately after their removal at autopsy and were kept at − 80°C until use. Samples were obtained from eight normal control subjects (mean age 46.6 years; range 26–78 years; post-mortem delay 16.0 ± 2.5 h) and eight ALS patients (mean age 65.5 years; range 24–79 years; post-mortem delay 7.3 ± 1.1 h; Table 1). Case A8 was previously reported as a case of sporadic juvenile ALS (Aizawa et al. 2000). Written informed consent was obtained from all subjects. The Ethics Committee of the University of Tokyo approved the experimental procedures used.
Table 1. Profiles of the cases analyzed in this study
Duration of illness (years)
Post-mortem delay (h)
Age, age of patient at death; C1–8, normal controls; A1–8, cases with ALS; Classic type, cases with upper and lower motoneuron signs in bulbar and two spinal regions.
Progressive bulbar type
Collection of tissue samples
Tissue samples from the frozen cerebellum (Cbl), cerebral cortex (Cx), and white matter (Wx) of eight control cases, which weighed under 10 mg, were collected into 1 mL of TRIZOL Reagent (Invitrogen Corp., Carlsbad, CA, USA). Frozen spinal cord samples of ALS and control cases were cut axially into 50-µm slices with a cryostat (Model HM500 O; MICROM, Walldorf, Germany). Samples were then attached to the slide glass followed by dissection into four parts [ventral gray matter (VG), dorsal gray matter (DG), ventral funiculus (VF), dorsal funiculus (DF)] in a freezing chamber (Model LD-550; TOMY, Tokyo, Japan) with a surgical binocular optical instrument and collected into 1 mL of TRIZOL Reagent. All spinal cord samples were derived from the upper thoracic level. The cerebellar molecular layer (ML) was collected by dissection in the same manner. These samples were kept at − 20°C until use.
Single cell isolation
Single cell isolation with an excimer laser microdissection system (Hamamatsu Photonics Ltd, Shizuoka, Japan) was previously described (Hashida et al. 2001). In brief, frozen sections of 20-µm thickness were sliced with the cryostat. After being attached to 1-mm-thick slide glasses made of artificial quartz and fixed with 100% methanol for 30 s, tissues were stained with 0.1% toluidine blue, and then dried completely. Single neurons were dissected at their margins with a focused laser beam and marginated neurons were collected in a test tube by a weak laser beam applied to the opposite side of the inverted quartz slide glass. One hundred cerebellar Purkinje cells (Purkinje), 200 cortical pyramidal neurons (Pyramidal), or 100 spinal motoneurons (Motor) were dissected from each section and collected into respective single test tubes that contained 200 µL of TRIZOL Reagent (Fig. 1). In some experiments, a single motoneuron of ALS and control cases was collected separately in a test tube by the same methods. As to the small neurons in the cerebellar granule cell layer (Granule) and substantia gelatinosa (Sensory) of the spinal cord, we dissected a part of the granular layer and substantia gelatinosa instead of dissecting individual neurons. All samples were kept at − 20°C until use.
RNA extraction and reverse transcription of tissue samples
From each tissue sample, total RNA was isolated with TRIZOL Reagent according to the manufacturer's instructions. One milligram of total RNA was incubated at 37°C for 15 min with Amplification Grade DNase I (Invitrogen) to remove contaminated DNA then heat-inactivated at 65°C for 10 min with 2.5 mm EDTA. First-strand cDNA was synthesized from the DNase I-treated total RNA with Ready-To-Go You-Prime First-Strand Beads (Amersham Biosciences Corp., Piscataway, NJ, USA) and 0.5 µm oligo(dT) primer, according to the manufacturer's instructions. The reverse transcription started with incubation at 25°C for 10 min, then at 37°C for 60 min, and was stopped by heating to 70°C for 10 min, which resulted in 33 µL of cDNA.
RNA extraction and reverse transcription of collected neurons
Forty microliters of chloroform were added to 200 µL of TRIZOL Reagent that contained collected neurons, then shaken vigorously for 1 min. It was centrifuged at 12 000 g for 30 min at 4°C, and then the supernatant was transferred to Phase Lock Gel Heavy (Eppendorf AG, Hamburg, Germany) with 120 µL of PCI (phenol/chloroform/isoamyl alcohol (25/24/1)). After vortex, it was centrifuged at 12 000 g for 15 min at 4°C. The supernatant was then transferred to a new tube that contained 10 µL of 3 m sodium acetate, 1 µL of a carrier mixture consisting of 0.95 µL of Ethachinmate (Nippon Gene Co., Tokyo, Japan), and 0.05 µL of Pellet Paint Co-precipitant NF (Novagen, Inc., Madison, WI, USA), which has no effect on fluorescent quantitative PCR. After we added 110 µL of 2-propanol to the tube and mixed by vortex, it was incubated at − 20°C for 2 h followed by centrifugation at 12 000 g for 30 min at 4°C. The pellet was rinsed with 75% ethanol twice and incubated at − 80°C overnight with 200 µL of 75% ethanol. After being rinsed twice again and air-dried completely, it was dissolved in 5 µL of the DNase reaction mixture, which contained 0.5 µL of 10 × DNase buffer, 0.2 units of DNase I, and 4.3 µL of sterile nuclease-free water. It was then incubated at 37°C for 15 min then denatured at 65°C for 10 min with 2.5 mm EDTA. DNase-treated samples were reverse transcribed with the Sensiscript RT Kit (QIAGEN GmbH, Hilden, Germany) according to the manufacturer's instructions with 0.5 mm oligo(dT) primer and 10 units of Prime RNase Inhibitor (Eppendorf AG) and incubated at 37°C for 120 min. The reaction was stopped by heating to 93°C for 5 min, which resulted in 20 µL of cDNA.
Standard preparation for quantitative PCR
In order to prepare an internal standard for quantitative PCR, gene-specific PCR products of more than 1 kb in length were amplified from human cerebellar cDNA with the following primers (amplified product lengths are also indicated): Primer set for GluR1, 1081 base pairs (bp): 5′-AGTTTGTGGTTCTCCCTGGGA-3′, 5′-ATCCTTCTTGTGGTCGCAGTG-3′; Primer set for GluR2, 1007 bp: 5′-AGCCCTCTGTGTTTGTGAGGA-3′, 5′-CAATAGTGGCTGCAGAAACAGG-3′; Primer set for GluR3, 1043 bp: 5′-TGGGTGCCTTTATGCAGCAA-3′, 5′-CGCGAGTCTCAGAGTGTTCCAT-3′; Primer set for GluR4, 1044 bp: 5′-TTGCCTACATTGGTGTCAGCG-3′, 5′-GGTCCGATGCAATGACAGC-3′; Primer set for β-actin, 1203 bp: 5′-CCCAGCCATGTACGTTGCTAT-3′, 5′-GCACGAAGGCTCATCATTCAA- 3′.
Each PCR reaction was run in 50 µL of a reaction mixture containing 200 µm of each primer, 1 mm dNTP Mix (Eppendorf AG), 5 µL of 10 × PCR buffer, and 1 µL of Advantage 2 Polymerase mix (BD Biosciences Clontech, Palo Alto, CA, USA). PCR amplification began with a 1-min denaturation step at 95°C, followed by 35 cycles of denaturation at 95°C for 10 s, annealing at 64°C for 30 s, and extension at 68°C for 80 s. PCR products were gel purified using Zymoclean Gel DNA Recovery Kit according to the manufacturer's protocol (Zymo Research, Orange, CA, USA). The absolute copy number of the PCR amplicons was then determined by measuring absorbance at 260 nm with a spectrophotometer (DU 650; Beckman Coulter, Fullerton, CA, USA). Serial dilutions ranging from 108 to 106, 104 to 103, and 102 to 101 copies per 2 µL of the DNA solution, which contained 100 ng of herring sperm DNA (Invitrogen) and Tris-EDTA buffer (pH 8.0), were then prepared.
Real-time quantitative PCR
PCR was performed using a LightCycler System (Roche Diagnostics, Mannheim, Germany). PCR primers and probes were designed from cDNA sequences for GluR1, GluR2, GluR3, GluR4, and β-actin (as an internal control), which were obtained from GeneBank (Table 2). To accomplish optimal conditions for PCR using oligo(dT) priming for first-strand synthesis and to avoid the flip/flop alternative splicing site, primers and probes for GluR1–GluR4 were designed to produce amplicons around 250 base pairs (bp) and to include the sequence between the putative fourth transmembrane domain (TM4) and the 3′ untranslated region (UTR; Steuerwald et al. 2000). Moreover, phosphate groups were attached to the 3′ end of LCRed640-dye containing probes to prevent probe extension.
Table 2. Sequences of primers and hybridization probes used for quantitative PCR
Amplified product length (bp)
bp, base pair; LCRed640, fluorescent dye for LightCycler; FITC, fluorescein isothiocyanate.
A set of standard and cDNA samples was amplified in duplicate in a master mixture (20 µL of total volume) that consisted of the following: 2 µL of 10 × FastStart Taq DNA polymerase (Roche Diagnostics) containing reaction mix, 4 mm MgCl2, 0.5 µm of each primer, 0.2 µm fluorescein isothiocyanate (FITC)-dye containing probes, 0.4 µm LCRed640-dye containing probes, and 1 unit of Uracil-DNA Glycosylase, heat-labile (UNG; Roche Diagnostics) to prevent PCR product carryover. Herring sperm DNA solution was also co-amplified as a negative control in each series of reactions.
The reactions started with 5-min of incubation at 30°C for activation of UNG, followed by denaturation of UNG and activation of FastStart Taq DNA polymerase at 95°C for 10 min. Templates were amplified by 45 cycles of denaturation at 95°C for 1 s and annealing of primers at 58°C for 10 s. This was followed by fluorescence acquisition and extension at 72°C for 12 s for GluR1–GluR4, and 9 s for β-actin.
The regional mRNA expression profile of GluR subunits in normal cases
We first investigated the regional expression level of each GluR subunit mRNA in the β-actin base. The amount of GluR2 mRNA in VG [24.1 ± 2.6 × 10−3 (mean ± SEM; n = 8)] was lower (from 2.5- to 12.3-fold) than in Cbl (296.2 ± 60.1 × 10−3), Cx (222.6 ± 25.3 × 10−3), ML (199.8 ± 22.3 × 10−3), or DG (60.7 ± 6.0 × 10−3; Mann–Whitney U-test, p < 0.001). The amount was the same as the level found in the white matter such as Wx (53.2 ± 9.2 × 10−3), VF (24.8 ± 2.6 × 10−3), and DF (19.8 ± 2.0 × 10−3; p > 0.01; Fig. 2a). The expression of GluR1, GluR3, and GluR4 mRNA was much lower than the expression of GluR2 mRNA and varied from region to region (Fig. 2b). The rank order of GluR3 > GluR4 > GluR1 was the common expression pattern in Wx, VG, DG, VF, and DF. However, the patterns of GluR1 > GluR4 > GluR3, GluR4 > GluR3 > GluR1, and GluR3 > GluR1 > GluR4 were found in Cbl, ML, and Cx, respectively. GluR1 mRNA was expressed in a very low level in VG and white matter (Wx, VF, and DF). GluR2 mRNA was predominant among GluR subunits in every region and shared from 83.7 ± 2.3% in VG to 91.9 ± 1.1% in Cx (Figs 2a–c). These values were statistically not different (Mann–Whitney U-test, p > 0.01).
The GluR subunits mRNA expression profile in neurons of normal cases
We next compared the copy number of GluR subunits mRNA between different neuronal subsets in the β-actin base. The amount of GluR2 mRNA was the least in motoneurons [52.9 ± 7.8 × 10−3 (mean ± SEM; n = 8)], which was lower (from 3.3- to 7.2-fold) than in Purkinje cells (381.0 ± 43.5 × 10−3), cortical pyramidal neurons (380.3 ± 90.8 × 10−3), sensory neurons in the substantia gelatinosa (191.6 ± 24.0 × 10−3), or cerebellar granule cells (176.3 ± 16.5 × 10−3; Mann–Whitney U-test, p < 0.001; Fig. 3a). The expression profile of GluR1, GluR3, and GluR4 mRNAs was different among neuronal subsets (Fig. 3b). GluR3 mRNA was the most abundant in Purkinje cells, while barely detectable in granule cells together with GluR1. GluR1 mRNA was expressed at a very low level in motoneurons. The results on tissue blocs (Figs 2a and b) and single neurons (Figs 3a and b) were consistent to the mRNA expression level and the profile of GluR subunits in the corresponding area. Thus, RNAs were not seriously damaged after dissection of neurons with the laser microdissector.
Although GluR2 shared the majority of GluRs subunits mRNA in all neuronal subsets as seen in the tissue level, the GluR2 mRNA expression in motoneurons [77.8 ± 2.0% (mean ± SEM; n = 8)] was lower than in other neuronal subsets (95.5 ± 0.5% in granule cells to 89.8 ± 1.2% in Purkinje cells; Mann–Whitney U-test, p < 0.01; Figs 3a–c). These results imply that AMPA receptor density, particularly of the GluR2-containing AMPA receptors, was the lowest in the motoneurons among neuronal subsets examined.
GluR subunits mRNA expression profile in the spinal cords from ALS and control cases
To determine whether selective reduction of GluR2 occurs in motoneurons of ALS cases, we first compared the copy number of GluR2 mRNA in the β-actin base in the spinal cords of ALS and control cases. Four parts (VG, DG, VF, DF) were dissected from the spinal cords of each case to clarify regional differences. Neither the expression level of GluR2 mRNA in ALS cases [VG, 15.3 ± 3.4 × 10−3; DG, 42.8 ± 6.4 × 10−3; VF, 23.4 ± 4.3 × 10−3; DF, 16.8 ± 1.7 × 10−3 (mean ± SEM; n = 8)] nor the proportion of GluR2 mRNA to total GluRs mRNA of ALS cases [VG, 91.4 ± 1.1% vs. 83.7 ± 2.3%; DG, 89.6 ± 1.5% vs. 85.8 ± 1.7%; VF, 89.5 ± 1.7% vs. 86.3 ± 1.0%; DF, 87.6 ± 1.7% vs. 85.4 ± 1.2% (mean ± SEM; n = 8)] was significantly different from values found in control cases in each spinal region (Mann–Whitney U-test, p > 0.01; Figs 2a and 4a,b).
GluR subunits mRNA expression profile in the spinal motoneurons of ALS and control cases
To test the possibility that selective reduction of GluR2 occurs in a subset of motoneurons of ALS cases, we compared the copy number of GluRs mRNA in the β-actin base in tissues collected from 100 spinal motoneurons from each individual case in the ALS and control groups. There was no significant difference between ALS and controls as to the copy number of GluR2 mRNA in motoneurons [ALS, 81.6 ± 12.8 × 10−3 (mean ± SEM, n = 8); control, 52.9 ± 7.9 × 10−3 (n = 8); Mann–Whitney U-test, p > 0.01]. There also was no significant difference in the proportion of GluR2 mRNA to total GluRs mRNA [ALS, 80.7 ± 1.8% (mean ± SEM, n = 8); control, 77.8 ± 2.0% (n = 8); Mann–Whitney U-test, p > 0.01].
In order to test the possibility that selective reduction of GluR2 occurs in a small proportion of motoneurons of ALS cases, we investigated the copy number of GluR2 mRNA at the single cell level. GluR2 and β-actin mRNAs of 91 single motoneurons of ALS cases (37 cells from A1, 27 cells from each of A2 and A5) and 81 single motoneurons of control cases (27 cells from each of C1, C3, and C5) were quantified. β-Actin mRNA was detected in 76 cells (83.5% of total cells) of ALS cases and 67 cells (82.7% of total cells) of control cases. Among β-actin mRNA-positive neurons, GluR2 mRNA was detectable in 44 cells (57.9%) of ALS cases and 36 cells (53.7%) of control cases. The frequency of GluR2 or β-actin mRNA positivity was not significantly different between the two groups (Mann–Whitney U-test, p > 0.01). Moreover, the amount of GluR2 mRNA in the β-actin base was not significantly different between ALS cases [126.9 ± 12.0 × 10−3 (mean ± SEM, n = 44)] and control cases [110.1 ± 13.4 × 10−3 (n = 36); Mann–Whitney U-test, p > 0.01; Fig. 5]. GluR2 mRNA was not detected in β-actin mRNA-negative neurons.
Results of quantitative measurements of mRNA by single cell RT–PCR on cultured rat spinal motoneurons demonstrated highly variable proportions of GluR2 mRNA to total AMPA receptor subunits mRNAs (from 2.1 to 63%) among various research groups (Greig et al. 2000; Vandenberghe et al. 2000b; Dai et al. 2001). With similar methods, the expression profile of AMPA receptor subunit mRNAs was quantitatively analyzed in the following cultured neurons: Purkinje cells (Lambolez et al. 1992; Brorson et al. 1999), cerebellar granule cells (Lambolez et al. 1992), hippocampal neurons (Jonas et al. 1994; Geiger et al. 1995; Lambolez et al. 1996; Tsuzuki et al. 2001), and spinal dorsal horn neurons (Vandenberghe et al. 2000b). The proportion of GluR2 mRNA to total AMPA subunit mRNAs, however, has varied greatly among neuronal subsets. Also, the strong GluR2 predominance that we found uniformly among tissues and neuronal subsets (from 89.8 to 95.5%) was not the rule. The variability of the results might be partially due to the difference in species (human or rat), methods of investigation (in situ or in vitro), culture length, and/or post-mortem delay. Above all, these researchers used common primer sets that were designed to recognize all four GluR subunits. However, primer sets were not exactly identical to the sequence of respective GluR subunits (Lambolez et al. 1992). This mismatch of primers to the template sequence might have influenced the efficiency of amplification in cases of small amounts of RNA, such as the substrate derived from single cells. In fact, GluR2 predominance (76%) among AMPA receptor subunits was also reported in situ in the rat hippocampus using serial analysis of gene expression (SAGE; Datson et al. 2001), which is one of the most reliable quantitative methods for global analyses of mRNA expression.
The expression level of each AMPA receptor subunit mRNA in spinal motoneurons was the lowest among the neuronal subsets examined and only twofold higher than in white matters (VF and DF; Figs 2a and 3b). The present results of mRNA expression do not necessarily reflect low protein expression and low AMPA receptor density of spinal motoneurons. Indeed, patch-clamp electrophysiological measurements on dissociated rat spinal cord cultures demonstrated a two- to threefold higher AMPA receptor current density in motoneurons than in dorsal horn neurons (Vandenberghe et al. 2000a), reflecting relatively high AMPA receptor density in the rat spinal motoneurons. To resolve the above inconsistency, analysis of AMPA receptor subunits at a protein level is essential, but previous studies with immunohistochemistry either failed to detect GluR2 in human motoneurons (Shaw et al. 1999) or reported inconsistent results in rat and mouse motoneurons (Morrison et al. 1998; Bar-Peled et al. 1999; Laslo et al. 2001).
As to other AMPA receptor subunits, their expression profiles were different from each other among neuronal subsets but consistent with results in animal brains. Cerebellar granule cells expressed low GluR1 and GluR3 mRNA (Keinänen et al. 1990; Monyer et al. 1991; Burnashev et al. 1992). Neurons in the substantia gelatinosa expressed high GluR2 mRNA (Furuyama et al. 1993; Tölle et al. 1993), and motoneurons expressed low GluR1 mRNA (Furuyama et al. 1993; Tölle et al. 1993). Rapid progress in the field of AMPA receptor trafficking has demonstrated the elaborate machinery that differentially controls activity-dependent trafficking of GluR1/GluR2 from the constitutive replacement of GluR2/GluR3 (Passafaro et al. 2001; Shi et al. 2001). Thus, different subunit expression profiles among neuronal subsets may reflect a balance between activity-dependent and constitutive AMPA receptor regulatory mechanisms.
Present analyses on both 100 motoneurons and single motoneurons indicate that selective GluR2 reduction (GluR2 hypothesis) does not occur and cannot be the mechanism underlying AMPA receptor-mediated neuronal death in ALS. Indeed, studies on cultured rat spinal cord neurons have suggested that the key determinant of selective AMPA receptor-mediated vulnerability of spinal motoneurons is the higher AMPA receptor current density in motoneurons than in dorsal horn neurons (Vandenberghe et al. 2000a), rather than the differential subunit composition of AMPA receptors between the two neuronal subsets. If this evidence in animal spinal cords is applicable to the human spinal cord, selective neuronal death of motoneurons in ALS is not the natural extension of preferential vulnerability of motoneurons to AMPA receptor-mediated neurotoxicity.
An alternative explanation for the selective motoneuronal death in ALS is the deficient GluR2Q/R site editing that we have demonstrated in the ventral gray matter of the ALS spinal cord (Takuma et al. 1999). When GluR2 mRNA is not edited at the Q/R site, even in a small proportion(s), the single channel conductance of AMPA receptors is markedly increased as a whole. Animals deficient of GluR2Q/R site editing die young due to neuronal death that could be rescued by GluR2(R) (Brusa et al. 1995; Higuchi et al. 2000). Also, Ca2+-permeable artificial GluR2 transgenic mice displayed motor neuron disease after 12 months of life (Feldmeyer et al. 1999). Greger et al. (2002) recently demonstrated that trafficking of GluR2 from the endoplasmic reticulum (ER) to the cell surface is dependent on whether the amino acid residue at the Q/R site is arginin (R) or glutamine (Q). Therefore, GluR2(R) was retained in the ER but GluR2(Q) and GluR1(Q at the Q/R site) were readily transported to the membrane. Effects of this GluR2 mRNA editing deficiency may be more conspicuous in the spinal motoneurons with relative low abundance of GluR2 mRNA and AMPA receptor subunit mRNAs. According to this scheme, a small increase of GluR2(Q) with deficient GluR2 RNA editing increases the proportion of GluR2(Q)-containing, and therefore Ca2+-permeable, functional AMPA receptors – thereby promoting excitotoxicity in ALS motoneurons.
This investigation was supported in part by a grant from The ALS Association, and a grant-in-aid for Scientific Research on Priority Areas from the Ministry of Education, Culture, Sports, Science and Technology of Japan (13210031,14017020).