Cortical and subcortical activity of glyceraldehyde-3-phosphate dehydrogenase (GAPDH) in postmortem brain of neurologically normal subjects. Values represent mean±SE of 5 subjects (age, 40±3 years). Numbers indicate Brodmann cerebral cortical areas 10 (frontal cortex), 21 (temporal cortex), 17 (occipital cortex), and 7b (parietal cortex); CC, cerebellar cortex; WM, white matter taken dorsal to head of the caudate; PUT, putamen; PI and PE, internal and external globus pallidus, respectively; and NL, nucleus thalamus lateralis.
Influence of age on activity of glyceraldehyde-3-phosphate dehydrogenase (GAPDH) in postmortem occipital cortex of 60 neurologically normal subjects aged from 1 day to 92 years.
Kish SJ, Lopes-Cendes I, Guttman M, Furukawa Y, Pandolfo M, Rouleau GA, Ross BM, Nance M, Schut L, Ang L, DiStefano L. Brain Glyceraldehyde-3-Phosphate Dehydrogenase Activity in Human Trinucleotide Repeat Disorders. Arch Neurol. 1998;55(10):1299–1304. doi:10.1001/archneur.55.10.1299
Although the abnormal gene products responsible for several hereditary neurodegenerative disorders caused by repeat CAG trinucleotides have been identified, the mechanism by which the proteins containing the expanded polyglutamine domains cause cell death is unknown. The observation that several of the mutant proteins interact in vitro with the key glycolytic enzyme glyceraldehyde-3-phosphate dehydrogenase (GAPDH) suggests that interaction between the different gene products and GAPDH might damage brain neurons.
To measure the activity of GAPDH in postmortem brain of patients with CAG repeat disorders.
Patients and Methods
Activity of GAPDH was measured in morphologically affected and unaffected brain areas of patients with 4 different CAG repeat disorders (Huntington disease, spinocerebellar ataxia 1 [SCA1], SCA2, and SCA3–Machado-Joseph disease), in brains of patients with Friedreich ataxia (a GAA repeat disorder) and Alzheimer disease, and in brains of matched control subjects.
Brain GAPDH activity was normal in all groups with the exception of a slight but statistically significant region-specific reduction in the patients with Huntington disease (caudate nucleus, −12%) and Alzheimer disease (temporal cortex, −19%).
The presence of the polyglutamine-containing proteins in CAG repeat disorders does not result in substantial irreversible inactivation or in increased activity of GAPDH in human brain.
TO DATE, at least 7 autosomal-dominant neurodegenerative disorders are known to be caused by CAG expansions in the coding regions of the genes. These include spinobulbar muscular atrophy, Huntington disease (HD), spinocerebellar ataxia type 1 (SCA1), SCA2, SCA3–Machado-Joseph disease (MJD), SCA6, SCA7, and dentatorubropallidoluysian (DRPLA) atrophy.1,2 The mechanism by which the proteins containing the expanded polyglutamine domains cause cell death is unknown. Recently, much attention has been devoted to the possibility that the abnormal proteins might cause death of neurons through protein-protein interactions, with the disease-specific regional specificity conferred by selective vulnerability of different cells or, perhaps, a specific brain-regional pattern of the interacting protein. In this regard, the protein products of the genes for HD (huntingtin), SCA1 (ataxin-1), spinobulbar muscular atrophy (androgen receptor), and DRPLA (atrophin) interact, at least in vitro, with the enzyme glyceraldehyde-3-phosphate dehydrogenase (GAPDH).3,4 This enzyme is a glycolytic enzyme responsible for the conversion of D-glyceraldehyde-3-phosphate to 1,3-diphosphoglycerate. As the in vitro data indicate that the binding between at least 1 of the mutant proteins (ataxin 1) and GAPDH is unexpectedly strong for a protein-protein interaction, being dissociated only by high (>1-mol/L sodium chloride) salt washes, the possibility exists that the abnormal proteins might cause irreversible degradation of the enzyme.4
As a key enzyme of glycolysis, GAPDH plays an important role in energy metabolism. In principle, therefore, the interaction between GAPDH and the polyglutamine-containing proteins might lead to decreased energy stores, which could trigger a degenerative process in susceptible cells. Interestingly, however, a separate body of experimental data suggests a competing scenario, namely, that overexpression of GAPDH might play a role in cell death mechanisms. In this regard, cerebellar granule cells undergoing apoptotic cell death show increased protein levels of GAPDH, with the neuronal death inhibited by treatment with a GAPDH antisense oligonucleotide.5- 7 As increased protein levels of GAPDH do not appear to be associated with elevated enzyme activity in the in vitro cell death model,7 a nonglycolytic role of GAPDH might be involved in the neuronal apoptotic process.8
As a first step toward understanding the possible involvement of GAPDH in CAG repeat disorders, we measured the total activity of the enzyme in a sampling of cortical and subcortical brain areas of normal subjects to determine whether the susceptibility of different brain areas to neurodegeneration in the various CAG repeat disorders might be related to regional differences in enzyme activity. To assess the possible influence of disease state on enzyme levels, we then measured the total activity of the enzyme in homogenates of degenerating and morphologically normal brain areas of patients with 4 different neurodegenerative disorders caused by expansions of CAG repeats, namely HD, SCA1, SCA2, and SCA3-MJD, and in matched control subjects. As disease controls, we included a group of patients with Friedreich ataxia (FA), an autosomal recessive disorder caused by an intronic GAA triplet expansion,9 and a group of patients with Alzheimer disease (AD), a disorder in which decreased brain GAPDH activity has previously been reported.10
Brains were obtained from groups of patients with the 4 CAG repeat and 2 non-CAG repeat neurodegenerative disorders. At autopsy, half of each brain and, for most of the patients with SCA, half of each spinal cord was fixed in formalin for neuropathological analysis, whereas the other half was frozen at −80°C until neurochemical analysis. The diagnosis in patients with SCA1, SCA2, SCA3-MJD, and HD was based on presence of characteristic clinical features (for SCA, imbalance, limb ataxia, dysarthria, and dysphagia; for HD, involuntary movements, including chorea), positive family history indicating autosomal dominant inheritance, and, at autopsy, neuropathological evidence of moderate to severe cell loss in the spinocerebellar system (for SCA, cerebellum, lower brainstem, and spinal cord) or basal ganglia (for HD, caudate, putamen, and globus pallidus). Polymerase chain reaction analysis of genomic DNA isolated from the brains confirmed the expanded CAG repeat size in the affected allele in all of the patients (range, SCA1, 54-72; SCA2, 39-44; SCA3-MJD, 69-80; HD, 39-59). All patients with FA fulfilled the diagnostic criteria for FA,11 including onset of ataxia and deep tendon areflexia in lower limbs before the age of 17 years and severe degeneration of the posterior columns, lateral corticospinal tract, and posterior spinocerebellar tract. The PCR analysis of genomic DNA from the brains of the patients with FA confirmed the expanded GAA repeats in all samples. All patients with AD had clinical evidence of dementia and, at neuropathological examination, an abundance of neuritic plaques and neurofibrillary tangles in neocortex and hippocampus in the absence of any other degenerative process. All of the neurological diseases were end-stage in their courses. Some neuropathological information on some of the patients with SCA1,12 FA,13 and AD14 has been published previously. Brain was also obtained at autopsy from 60 control subjects aged 1 day to 92 years, who died without evidence of neurological or psychiatric disease or brain neuropathological abnormality. Table 1 and Table 2 show the ages, sex, and suspected cause of death for the control and patient subjects, respectively, with the number of CAG repeats for the patients with HD and SCA shown in Table 2.
For the neurochemical analyses, cerebral cortical subdivisions were excised according to Brodmann classification, with the caudate nucleus (intermediate portion of slice 4) and nucleus lateralis of the thalamus dissected as described by Kish et al15 and in the atlas of Riley,16 respectively.
The GAPDH activity (measured in the forward reaction) was determined using minor modifications of a spectrophotometric procedure,17 which used brain homogenates that had been sonicated (2×15 strokes) in 50-mmol/L tris(hydroxymethyl)aminoethane phosphate (pH 7.5), 0.32-mol/L sucrose, 1.0-mmol/L sodium EDTA acid, 1.0-mmol/L dithiothreitol, and 0.5% Triton X-100. The incubation mixture contained 135-mmol/L tris(hydroxymethyl)aminoethane acetate, 0.14-mmol/L oxidized nicotinamide-adenine dinucleotide (NAD+), 17-mmol/L disodium arsenate, 3.3-mmol/L cysteine, 0.25 mg of tissue wet weight, and 1.5-mmol/L glyceraldehyde-3-phosphate in a total volume of 3.0 mL. The assay conditions were optimized for human brain to ensure that the concentrations of glyceraldehyde-3-phosphate and NAD+ were at maximally stimulating levels. Changes in absorbance at 340 nm, corresponding to the reduction of NAD+, were determined using a spectrophotometer (Hitachi model U-2000, Tokyo, Japan) at 30°C. Boiled tissue homogenates, which gave values identical to those of samples employing buffer in place of glyceraldehyde-3-phosphate or homogenate, were used as blanks.
To establish whether activity of GAPDH was heterogeneously distributed among different brain areas, a 1-way analysis of variance (ANOVA) at the .05 criterion level was employed. To determine the relationship between age and postmortem interval (interval between death and freezing of the brain) on enzyme activity, Pearson correlation coefficients were computed. Possible differences between brain GAPDH activity in patient vs control groups were assessed using 1-way ANOVA followed by the Tukey test if ANOVA was significant at .05 criterion when comparisons were made between a control group and multiple disease conditions in which the same brain areas were examined (controls vs patients with SCA1, SCA2, SCA3-MJD, and FA). For other comparisons involving different brain areas between a control group and a single disease state (HD and AD), the Student 2-tailed t test was employed. The null hypothesis was that there would be no effect of disease condition on enzyme activity.
The GAPDH activity in the human brain was linear with respect to enzyme protein level and time across the range used. As expected,18 enzyme activity was inhibited by iodoacetic acid (50% inhibition at approximately 150 µmol/L; n=3, parietal cortex). The Km (Michaelis-Menten constant) (mean of 3 determinations in control parietal cortex) for glyceraldehyde-3-phosphate and NAD+ were 101 and 6.4 µmol/L, respectively. Regression analyses revealed no statistically significant influence of postmortem interval on brain GAPDH activity for the control or patient groups (P>.05).
An ANOVA revealed a heterogeneous (P<.001) distribution of GAPDH activity among the different brain areas of normal controls (mean age, 40 years) (Figure 1). However, with the exception of the white matter (taken dorsal to the caudate nucleus), in which enzyme levels were modestly lower than in gray matter areas, activity of the enzyme was similar in all examined cortical and subcortical brain areas.
To examine the influence of age on GAPDH activity, enzyme levels were determined in occipital cortex of a total of 60 normal controls aged from 1 day to 92 years. As shown in Figure 2, substantial levels of GAPDH were present in occipital cortex during the perinatal period, with a moderate (approximately 60%) increase in activity to about 8 months of age. This was followed by a slight increase in activity throughout childhood and adulthood. Regression analysis revealed a statistically significant age-related increase in enzyme activity across the entire age range (Figure 2) and from 1 day to 8 months (r=0.83; P <.001; n=13).
As shown in Table 3, no statistically significant differences in brain GAPDH activity were observed between the patients with SCA1, SCA2, SCA3-MJD, and FA as compared with the control group. For the patients with HD, a slight but statistically significant reduction in enzyme activity was observed in caudate nucleus (−12%; P<.001), whereas in the AD group, GAPDH activity was decreased by 19% in the temporal cortex (P<.02). No statistically significant correlations (Spearman rank) were observed between number of CAG repeats and GAPDH activity in any of the 4 CAG repeat disorders in any brain region examined (P>.05).
Our major finding is that GAPDH activity is normal or near normal in brains of patients with 4 different CAG repeat disorders.
To determine whether the brain regional pattern of cell loss in different trinucleotide repeat disorders might be related to differences in amount of GAPDH, we examined the regional distribution of GAPDH activity in cortical and subcortical areas of the normal human brain. However, activity of GAPDH was similar in all gray matter brain regions examined, including regions affected (for HD, putamen and globus pallidus; for SCA disorders, cerebellar cortex) and relatively unaffected (cerebral cortex, thalamus) by neuronal loss in the CAG repeat disorders. Our human brain data are consistent with those of Mizuno and Ohta,19 who reported a lack of heterogeneity of GAPDH activity among different brain areas of the rat and little difference between brain enzyme activity in adult vs aged rats. These findings in mammalian brain suggest that susceptibility of different brain areas to neuronal degeneration is unlikely to be explained by regional differences in total GAPDH levels.
We found that activity of GAPDH, measured under conditions that would assess maximal amount of the active enzyme, was normal in the brain of patients with the 4 different CAG repeat disorders, with the exception of a slight (12%) reduction, in the HD group, of enzyme activity in the caudate nucleus, a brain region affected by severe neuronal loss. In the case of HD, our finding of only slightly decreased brain GAPDH activity is generally consistent with the results of a recent investigation in which normal levels of the enzyme were found in the brain (cortical and subcortical areas) of patients with HD (who had not undergone assessment for CAG repeat number).20 The difference between the results of this and our own studies might be due to different methods of GAPDH measurement (enzyme activity restricted to subcellular fraction assessed by reverse enzyme reaction in the other study20; total enzyme activity in brain homogenates assessed by forward reaction in our investigation) or due to differences in patient severity of illness in both studies. In agreement with an early investigation,10 activity of GAPDH was slightly (−19%) reduced in the temporal cortex of the patients with AD. The modest reduction of brain GAPDH activity restricted to abnormal areas in HD and AD is consistent with the possibility that GAPDH might be irreversibly damaged in some disease-specific neurodegenerative processes (eg, by associated oxidative processes21,22) or that the enzyme has some slight preferential localization to neurons degenerated in both conditions.
Two scenarios have been suggested regarding the possible nature of the involvement of GAPDH and the CAG expansion. First, the interaction between polyglutamine-containing proteins and GAPDH results in reduced activity of this energy-metabolizing enzyme and, consequently, cell death in susceptible brain areas due to decreased energy stores.3,4 Second, the abnormal protein-GAPDH interaction leads to overexpression of the enzyme and consequent cell death by apoptosis,5- 8 a form of cell death reported to occur in brain of patients with HD23- 25 and SCA1.26 The abnormal gene products in SCA1, SCA3-MJD, and HD are widely distributed throughout the brain, being present in normal and degenerating brain regions.27- 29 However, we found that activity of GAPDH, which can bind, at least in vitro, to several of these mutant proteins,3,4 was normal or near normal in morphologically affected and unaffected brain regions in the 4 CAG repeat disorders. Thus, we conclude, on the basis of our postmortem brain findings, that the presence of the mutant proteins in CAG trinucleotide repeat disorders does not result in any substantial, irreversible loss of GAPDH protein or overexpression of GAPDH (as inferred by activity levels measured under maximal substrate conditions) consequent to the polyglutamine protein-GAPDH interaction. We emphasize, however, that these data do not at all exclude the possibility that glutamine-repeat proteins might reversibly inhibit GAPDH activity in vivo and thereby lead to cell death, or that the abnormal proteins could result in functionally significant changes in protein levels of the enzyme, perhaps in specific subcellular or neuronal compartments, which are not reflected in altered total activity of GAPDH. The molecular mechanism by which CAG gene expansions cause neurodegenerative disease requires further investigation.
Accepted for publication February 25, 1998.
This study was supported by grant 26034 from the National Institute of Neurological Disorders and Stroke, National Institutes of Health, Bethesda, Md (Dr Kish).
Some postmortem brain tissue was received from the Tissue Donation Program of the National Ataxia Foundation, Wayzata, Minn, and the Canadian Brain Tissue Bank, Toronto, Ontario.
Reprints: Stephen J. Kish, PhD, Human Neurochemical Pathology Laboratory, Clarke Institute of Psychiatry, 250 College St, Toronto, Ontario, Canada M5T 1R8 (e-mail: firstname.lastname@example.org).