Region- and cell-specific neuronal loss is a major hallmark of Alzheimer’s disease (AD) brains. The pattern of damage is not distributed uniformly, so that distinct regions of the AD brain can be classified as either affected or spared. The mechanism responsible for the regional specificity of pathology has yet to be resolved. Cell loss is particularly severe in glutamatergic neurones of the hippocampus and neocortex. Glutamate is the major mediator of excitatory signals in the mammalian brain. Uptake of released transmitter by the major glial glutamate transporter, excitatory amino acid transporter 2 (EAAT2), is necessary for terminating glutamate transmission, preventing excitotoxicity by over-excitation of neurones, and inhibiting cross-talk from neighbouring neurones.
The focus of this project was to determine if quantitative or qualitative changes in EAAT2 might provide a basis for the regional loss of neurones in the AD brain. To this end, a targeted, discovery-based approach was utilized to determine the sequence of EAAT2 splice variants present in the adult human brain. Several known and two previously unknown transcripts were cloned and sequenced. The known variants cloned from human brain were the C-terminal variant, EAAT2b, and a variant lacking exon 7, EAAT2Δ7. Two novel transcripts were found that lack both exon 7 and 9 with either the wild-type C-terminus, EAAT2Δ7Δ9, or the known alternative C-terminus, EAAT2bΔ7Δ9. This represents the first time exon-skipping splicing events have been demonstrated in EAAT2b.
To determine the proportion of these variants in normal and disease states, a Real Time RT-PCR assay using SYBR® green technology was developed. EAAT2, EAAT2b, EAAT2Δ7, and EAAT2Δ9 were measured in four areas of 37 human brains (15 control, 12 AD, and 10 AD with Lewy body dementia). Pilot assays of the double exon-skipping variants showed very low expression, therefore they were not included in the broader study. Four brain regions were chosen according to their pathological relevance to AD. Susceptible regions included the inferior frontal cortex and inferior temporal cortex; relatively spared regions included the posterior motor and occipital cortices. Overall, there was a significant reduction in total EAAT2 mRNA expression for both neurodegenerative disease groups compared with controls. The main finding of this study was that lower expression of EAAT2 message in AD cases occurred in both affected and spared regions. This was contrary to the hypothesis proposed. This indicates that loss of EAAT2 is not the primary cause of the regional loss of neurones in AD. However, splice variant expression displayed regional selectivity, with lower levels of exon-skipping variants in spared regions, higher expression of exon-skipping variants in affected regions, and the loss of the functional variant in the inferior temporal cortex of AD cases. Changes in the relative expression of exon-skipping EAAT2 variants indicate that the amount of functional glutamate transport may be regulated in a complex manner by the presence of alternative EAAT2 mRNAs that do not bind glutamate. Division of neurodegenerative disease cases by region based on pathological scoring demonstrated an increase in EAAT2 expression with increasing pathological severity, suggesting that the relative increase of EAAT2 mRNA in AD susceptible regions is a neuroprotective response to the dispersal of glutamate that occurs during neuronal loss. Further analysis detected variations in the pattern of exon-skipping splice variant expression according to APOE genotype and neurodegenerative disease type. This suggests that genetic susceptibility can influence glutamate transport and/or expression and may delineate a subtype of AD. Overall, the Real Time RT-PCR study indicated that alternative splicing is an endogenous mechanism of regulation, and that each variant can be individually controlled in a regional and disease-specific manner.
To determine the consequences of differences in exon-skipping transcript levels, EAAT2 variants were expressed in Xenopus lævis oöcytes. Two-electrode voltage clamp studies revealed a negative effect of EAAT2Δ7, EAAT2Δ9, and EAAT2Δ7Δ9 on wild-type glutamate uptake. This effect varied in strength and in the shapes of the dose-response curves. This may be due to the different functional residues that each variant lacks. The ability of mRNA splice variants to alter protein trafficking and/or function represent an emerging mechanism for regulating glutamate uptake. In contrast to EAAT2Δ7, EAAT2Δ9 was able to form homomultimers and insert into the plasma membrane of Xenopus lævis oöcytes, indicating that the ability to form multimers is an important signal for forward trafficking from the ER.
A discovery-based proteomics project was undertaken to search for the protein variants of EAAT2 that occur in human brain. This produced many proteins of interest that were EAAT2-immunoreactive and varied in size and isoelectric charge, but this could not be verified by peptide mass fingerprinting due to technical considerations.