Numerous studies support a gain of toxic function of Aβ due to post-translational modifications such as isoaspartate formation, nitrosylation , and formation of pyroglutamate [10, 11, 30, 32]. In order to assess the pathological consequence of enhanced pE-Aβ in vivo, several tg mouse models expressing truncated human Aβ with an N-terminal Q to E substitution have been generated [17–19]. Although these models display differences in terms of behavior and neuronal loss, they commonly suggested that an increase of pE-Aβ enhances neuronal dysfunction. With regard to TBA2 and TBA2.1 mouse models, the ETNA gene product is the same. However, pE3-Aβ formation in ETNA is not accelerated by an exchange of E by Q at position 3 of Aβ. As shown in previous studies, the N-terminal E residues represent worse substrates for QC (and isoQC) compared to Q, and E residues should not cyclize spontaneously considering a half-life of 10–40 years in vivo[2, 3]. In accordance with such slower formation of pE-Aβ(3–42) also in vivo, we observed the behavioral changes and neuronal loss at a later time point compared to TBA2.1.
Furthermore, in ETNA the tg product, Aβ(3–42), is mainly expressed in the lateral striatum compared to the hippocampal and brainstem expression in TBA2.1. In this brain region ETNA animals show progressive intraneuronal pE3-Aβ formation associated with neuroinflammatory, apoptotic, and neurodegenerative processes, including but not limited to a reduced DARPP-32 immunoreactivity. The behavioral phenotype of single tg ETNA mice, hyperactivity and impaired acoustic sensorimotor gating, is associated with striatum the brain region with distinct neuropathology.
Key feature of the neuropathological process is a progressive site-specific expression of pE3-Aβ, which is associated with increased caspase 3 and GFAP expression, ultimately resulting in significant cell loss of medium-sized spiny neurons (MSN), the major cell type in the striatum .
Interestingly, TBA2.1 mice have also shown expression of the Aβ(Q3-42) construct in the striatum and hippocampus. Only in the latter brain region, pE3-Aβ formation and neurodegeneration is observed in these animals . ETNA mice, which use the same Thy1.2 promoter and expression cassette and only differ in one amino acid, express the Aβ(E3-42) construct in the same regions, but surprisingly pE3-Aβ formation and neurodegeneration mainly occurs in the striatum. In both animal models, TBA2.1 and ETNA, neuronal loss is exclusively observed in vicinity of pE3-Aβ, arguing for pE3-Aβ as the neurotoxic Aβ species.
Compared to other unmodified Aβ, pE3-Aβ had been shown to accelerate accumulation of several Aβ species, serving as seed of Aβ aggregation [15–17]. Although, the presence of other Aβ species might contribute for the phenotype of ETNA animals, the properties of pE3-Aβ and the spatio-temporal correlation with neuropathology argue for pE3-Aβ as crucial factor. In summary, the presence of pE3-Aβ seems to trigger the coaggregation and to influence the neurotoxic properties of other Aβ species.
Despite for the generation of TBA mice the same Thy1.2 promoter was used as for ETNA mice, in all models a different brain region showed the most expression of the tg construct. In TBA2 the most tg expression was found in the cerebellum , TBA2.1 mainly expressed Aβ in the hippocampus , and ETNA mice showed most expression in the striatum. The different insertion sites of the tg construct into the genome might have led to the regional variation of expression intensity, but the exact mechanism remains unclear.
Also, the kinetics of pE3-42 deposition and neuropathology implies a crucial role of the modified peptide for neuronal loss. Already at the age of 1 month (first time point of analysis), hom E8 show expression of the truncated Aβ(3–42) construct in the lateral striatum and severe Aβ protein levels of brain homogenates were detected by ELISA. With 3 months of age Aβ expression of hom E8 animals is increasing, reaching a plateau at 6 months. At this age, immunohistochemical stainings then revealed exponential pE3-Aβ formation being associated with immunoreactivity specific for activated astroglia, increased caspase 3 activity, decrease of DARPP-32 expression, and blunted NeuN immunoreactivity, which altogether describe neuroinflammatory and apoptotic neurodegenerative processes.
In temporal correlation of prominent pE3-Aβ formation first behavioral alterations were detected in single tg ETNA animals and at the age of 6 months 90% of hom E8 animals showed decreased reactivity on a given acoustic signal. These animals showed obvious neuronal loss in the lateral striatum at the age of 9 months and hyperactivity was detected at the age of 10 months.
In het/het E85 animals the same chronology of neuropathology and behavioral phenotype was observed with a delay of 2–4 months.
Moreover, overexpression of hQC has been described to enhance pE3-Aβ dependent pathology and behavior in 5XFAD-hQC mice . To further analyze the role of QC on pE3-Aβ formation and neuropathology, E8 and E85 animals were intercrossed with human QC overexpressing mice to generate double tg ETNA-hQC mice.
These mice showed elevated pE3-Aβ levels and increased neuropathology compared to single tg ETNA mice providing evidence for QC-dependent formation of neurotoxic pE3-Aβ in vivo. Overexpression of hQC had no influence on total Aβ levels detected by ELISA in both ETNA-hQC lines (E8-hQC and E85-hQC), but increased pE3-Aβ levels and numbers of pE3-Aβ positive cells in the lateral striatum. Double tg E8-hQC animals showed hQC gene dose-dependent increase of GFAP and decrease of DARPP-32 reactivity in the striatum at 6 months of age. Neuronal numbers were quantified in the basolateral striatum of these animals and revealed a significantly increased neuronal loss of 45% for hom/hom E8-hQC animals compared to 21% for hom/wt E8-hQC animals with hQC wt expression. Enhanced pE3-Aβ formation and pathology in hQC overexpressing animals provide direct evidence for a QC-driven process, limited by the available amount of QC.
The striatal pathology of ETNA contrasts with other mouse models such as TBA2.1 or hAPP tg mice [35–37]. In most of these models, hippocampal (and cortical) degeneration has been observed. ETNA mice display only weak expression of the construct in the pyramidal cell layer of the hippocampus (Additional file 1: Figure S1C), from the age of 3 months and no pE3-Aβ formation is observed up to the age of 9 months in E5, E8, and E85 animals (data not shown).
In accordance with the striatal neuronal loss, sensorimotor gating impairments, similar to striatal alterations in HD, have been observed in ETNA mice . Likewise, in mice with a genetic deletion of DARPP-32 or with point mutations in phosphorylation sites of DARPP-32, the effects of stimulating drugs on sensorimotor gating were strongly attenuated .
Impairment of acoustic sensorimotor gating is associated with striatal neurodegeneration occurring at the sites of pE3-Aβ expression and diminished DARPP-32 immunoreactivity.
Similarly, the home cage hyperactivity displayed by single tg ETNA mice is likely to be a consequence of the underlying striatal pathology, possibly caused by a loss of GABAergic inhibitory neurotransmission resulting in a disinhibition of motor function.
While hyperkinetic movement disorders represent an important subgroup of neurodegenerative disorders, not all of those affect the striatum . Among them, HD shows the highest incidence and has been associated with disturbed dopaminergic as well as downstream DARPP-32 neurotransmission [41, 42]. A loss of inhibitory GABAergic MSNs represents a key feature of the disease and is causal for hyperactivity . In addition to their relevance for pE3-Aβ toxicity in AD, ETNA lines might model selected aspects of HD.