Novel rat Alzheimer's disease models based on AAV-mediated gene transfer to selectively increase hippocampal Abeta levels

Abstract

Background: Alzheimer's disease (AD) is characterized by a decline in cognitive function and accumulation of amyloid-beta peptide (Abeta) in extracellular plaques. Mutations in amyloid precursor protein (APP) and presenilins alter APP metabolism resulting in accumulation of Abeta42, a peptide essential for the formation of amyloid deposits and proposed to initiate the cascade leading to AD. However, the role of Abeta40, the more prevalent Abeta peptide secreted by cells and a major component of cerebral Abeta deposits, is less clear. In this study, virally-mediated gene transfer was used to selectively increase hippocampal levels of human Abeta42 and Abeta40 in adult Wistar rats, allowing examination of the contribution of each to the cognitive deficits and pathology seen in AD.

Results: Adeno-associated viral (AAV) vectors encoding BRI-Abeta cDNAs were generated resulting in high-level hippocampal expression and secretion of the specific encoded Abeta peptide. As a comparison the effect of AAV-mediated overexpression of APPsw was also examined. Animals were tested for development of learning and memory deficits (open field, Morris water maze, passive avoidance, novel object recognition) three months after infusion of AAV. A range of impairments was found, with the most pronounced deficits observed in animals co-injected with both AAV-BRI-Abeta40 and AAV-BRI-Abeta42. Brain tissue was analyzed by ELISA and immunohistochemistry to quantify levels of detergent soluble and insoluble Abeta peptides. BRI-Abeta42 and the combination of BRI-Abeta40+42 overexpression resulted in elevated levels of detergent-insoluble Abeta. No significant increase in detergent-insoluble Abeta was seen in the rats expressing APPsw or BRI-Abeta40. No pathological features were noted in any rats, except the AAV-BRI-Abeta42 rats which showed focal, amorphous, Thioflavin-negative Abeta42 deposits.

Conclusion: The results show that AAV-mediated gene transfer is a valuable tool to model aspects of AD pathology in vivo, and demonstrate that whilst expression of Abeta42 alone is sufficient to initiate Abeta deposition, both Abeta40 and Abeta42 may contribute to cognitive deficits.

Figure 1

Transgene expression in rat hippocampus, 3 weeks post-infusion. A group of test animals (n = 3 per group) was injected with AAV1 vectors and brains analysed for transgene expression (anti-amyloid beta protein, Chemicon, MAB1560, 6E10, specific for human Aβ). No Aβ expression was observed in naïve rat brain. Infusion of AAV1 vectors expressing BRI-Aβ42 and BRI-Aβ40 fusions or APPsw resulted in extensive expression of amyloid β in the dentate gyrus and CA1 fields of the hippocampus, with distinctive expression patterns for each vector. Left panel: low magnification view of unilateral hippocampus, scale bar 250 μm. Middle panel: higher magnification view of dentate gyrus (DG), scale bar 50 μm. Right panel: higher magnification view of CA1, scale bar 50 μm.

Figure 2

Behavior in the open field. 12 rats per group were used for behavioral testing. A. BRI-Aβ40+42 animals crossed significantly more lines than naïve controls or any other treatment group (ANOVA P = 0.027; BRI-Aβ40+42 > naïve, **p < 0.01). B. Although overall there was no significant difference between treatment groups (ANOVA P = 0.24), BRI-Aβ40+42 animals spent the least amount of time in the center of the field.

Figure 3

Morris water maze. A. Repeated measures ANOVA on latency to find the platform during acquisition shows that overall the difference between treatment groups was not significant (P = 0.12). However, naïve animals took fewer trials to learn the location of the platform than AAV treatment groups, reaching a minimum latency by day 3. BRI-Aβ40+42 animals had a noticeably higher latency on day 3 of acquisition than other groups. B. BRI-Aβ40+42 animals had the highest mean latency over all 20 trials of all treatment groups (ANOVA P = 0.07). C. Repeated measures ANOVA on pathlength to find the platform during acquisition shows that overall the difference between treatment groups was not significant (P = 0.061). D. BRI-Aβ40+42 animals took a significantly longer pathlength to find the platform over all 20 trials (ANOVA P = 0.02; BRI-Aβ40+42 > naïve, ** p < 0.01). E, F. Animals were put through a probe trial 24 hr after the last training trial. BRI-Aβ40+42 animals showed a non-significant reduction in time (ANOVA P = 0.2) and distance (ANOVA P = 0.13) spent in the target quadrant.

Figure 4

A. Passive avoidance. There was no difference in baseline latencies between the treatment groups (ANOVA P = 0.97) but at 24 hr BRI-Aβ40 animals had a reduced latency to re-enter the dark chamber compared to naïve controls (ANOVA P = 0.03; BRI-Aβ40 < naive, *p < 0.05). B, C. Novel object recognition. At 3 months post-injection BRI-Aβ42, BRI-Aβ40+42 and APPsw animals spent less total time (O1+O2+O3+N) exploring than naïve controls (B) but still explored the novel object (N) more than the familiar object, O3 (C). Total exploration time: ANOVA P = 0.003; APPsw < naive, p < 0.01, BRI-Aβ42 < naïve, p < 0.05.

Figure 5

A. Characterization of Aβ levels in hippocampus by Aβ sandwich ELISA. 12 rats per treatment were used for biochemical and histological analyses. Naïve rats have very low levels of either Aβ species. AAV-BRI-Aβ40 injected animals display an increase in both RIPA- and FA-soluble Aβ40 at 3 months post-injection. There is a marked increase in FA-soluble (RIPA-insoluble) Aβ42 in AAV-BRI-Aβ42 animals at 3 months. B. Amyloid pathology in AAV-BRI-Aβ42 animals. Only BRI-Aβ42 animals showed diffuse plaque-like structures within the hippocampus following immunostaining with pan anti-Aβ 1–16 antibody. These deposits did not stain with thioflavin S or Congo Red (not shown). Magnification 100x.