Impaired Cognitive Flexibility With Preserved Learning in an Amyloid Precursor Protein Knock-In Mouse Model of Amyloidopathy

Abstract

Alzheimer's disease is a debilitating neurodegenerative condition characterized by amyloid beta plaques and tau neurofibrillary tangles, which leads to progressive cognitive decline. Several new mouse models of fast amyloid deposition have been generated with compound mutations, but how these affect high-level cognitive function is still not fully understood. Four cohorts of a second-generation amyloid precursor protein knock-in mouse model, App^{NL-G-F/NL-G-F}, which develops aggressive amyloidopathy, were compared with two different control groups that do not produce plaques (App^NL/NL and wildtype littermates), on touchscreen-based tests of learning and cognitive flexibility. App^{NL-G-F/NL-G-F} mice learned to discriminate between two visual stimuli during the pairwise visual discrimination (PVD) task but were impaired when the reward contingencies were reversed (the PVR task). Analyses of the correction trials indicated perseverative behavior. One cohort was further tested on the touchscreen Extinction test, which isolates the ability to withhold responding to a previously rewarded stimulus. The App^{NL-G-F/NL-G-F} mice extinguished their responding no differently than the App^NL/NL control group. These results indicate that compound mutations in App driving fast accumulation of plaques in this mouse model impair cognitive flexibility and may serve as a preclinical target for putative therapeutic drugs.

Keywords: APP; executive dysfunction; extinction; pairwise visual discrimination; reversal learning.

FIGURE 1

Female App ^{NL‐G‐F/NL‐G‐F} are impaired in reversal learning compared to both control groups. Pairwise visual discrimination learning and reversal percent correct responses, number of correction trials, and perseveration index for female mice. (A) Figure legend, cohorts, ages, and the images used during touchscreen testing. Trials to acquisition criterion, percent correct responses, correction trials, and perseveration Indices for (B–E) Cohort 1 (App ^{NL‐G‐F/NL‐G‐F} vs. App ^NL/NL), (F–I) Cohort 2, Timepoint 1 (App ^{NL‐G‐F/NL‐G‐F} vs. App ^{NL‐G‐F/NL}), (J–M) Cohort 2, Timepoint 2 (App ^{NL‐G‐F/NL‐G‐F} vs. App ^{NL‐G‐F/NL}). (N–Q) Cohort 3 (App ^{NL‐G‐F/NL‐G‐F} vs. Wt littermates), and (R–U) Cohort 4 (App ^{NL‐G‐F/NL‐G‐F} vs. Wt littermates). App ^{NL‐G‐F/NL‐G‐F} females required more sessions to acquire the task compared to App ^NL/NL controls (F). There were no significant differences in trials to acquisition in any other Cohort or timepoint. App ^{NL‐G‐F/NL‐G‐F} females were non‐significantly impaired compared with App ^NL/NL controls (p = 0.052) at PVR in Cohorts 1 (C) and 2 (Timepoint 1; G), with increased correction trials and perseveration in Cohort 2, Timepoint 1 (H, I). There were no significant differences between groups upon retesting at 12–13 months of age (Cohort 2, Timepoint 2; K–M). When compared with littermate controls, both Cohorts of App ^{NL‐G‐F/NL‐G‐F} female mice were impaired in reversal learning, with significant genotype effects in correct responses (O, S) and correction trials (P, T) in both Cohorts, a significant genotype effect in perseveration index in Cohort 4 (U), and a non‐significant trend toward a genotype effect in Cohort 3 (Q). Red symbols and lines depict App ^{NL‐G‐F/NL‐G‐F} groups, blue depict App ^NL/NL mice, yellow depict littermate wildtype controls. *p < 0.05, **p < 0.01, ***p < 0.001. Data depicted as mean ± SD (for PVD sessions to criterion) or standard error of the mean (SEM) (for baseline PVD and all PVR sessions). Cohort 1: App ^NL/NL n = 5, App ^{NL‐G‐F/NL‐G‐F} n = 7; Cohort 2 Timepoint 1: App ^NL/NL n = 15, App ^{NL‐G‐F/NL‐G‐F} n = 15; Cohort 2 Timepoint 2: App ^NL/NL n = 15, App ^{NL‐G‐F/NL‐G‐F} n = 10; Cohort 3: Wt littermates n = 3, App ^{NL‐G‐F/NL‐G‐F} n = 5; Cohort 4: Wt littermates n = 10, App ^{NL‐G‐F/NL‐G‐F} n = 12.

FIGURE 2

Male App ^{NL‐G‐F/NL‐G‐F} are impaired in reversal learning compared to both control groups. Pairwise visual discrimination learning and reversal percent correct responses, number of correction trials, and perseveration index for male mice. (A) Figure legend, cohorts, ages, and the images used during touchscreen testing. Trials to acquisition criterion, percent correct responses, correction trials, and perseveration indices for (B–E) Cohort 1 (App ^{NL‐G‐F/NL‐G‐F} vs. App ^{NL‐G‐F/NL}), (F–I) Cohort 2, Timepoint 1 (App ^{NL‐G‐F/NL‐G‐F} vs. App ^{NL‐G‐F/NL}), (J–M) Cohort 2, Timepoint 2 (App ^{NL‐G‐F/NL‐G‐F} vs. App ^{NL‐G‐F/NL}). (N–Q) Cohort 3 (App ^{NL‐G‐F/NL‐G‐F} vs. Wt littermates), and (R–U) Cohort 4 (App ^{NL‐G‐F/NL‐G‐F} vs. Wt littermates). There were no significant differences in trials to acquisition in any Cohort or timepoint. App ^{NL‐G‐F/NL‐G‐F} males were significantly impaired compared with App ^NL/NL controls at PVR in Cohorts 1 (C) and 2 (Timepoint 1; G), with increased correction trials and perseveration in Cohort 1 (D, E). There were no significant differences between groups upon retesting at 12–13 months of age (Cohort 2, Timepoint 2; K‐M). When compared with littermate controls, both Cohorts of App ^{NL‐G‐F/NL‐G‐F} mice were impaired in reversal learning (correct responses, correction trials, and perseveration index; O–Q, S–U). Red symbols and lines depict App ^{NL‐G‐F/NL‐G‐F} groups, blue depict App ^NL/NL mice, yellow depict littermate wildtype controls. *p < 0.05, **p < 0.01, ***p < 0.001. Data depicted as mean ± SD (for PVD sessions to criterion) or SEM (for baseline PVD and all PVR sessions). Cohort 1: App ^NL/NL n = 5, App ^{NL‐G‐F/NL‐G‐F} n = 11; Cohort 2 Timepoint 1: App ^NL/NL n = 15, App ^{NL‐G‐F/NL‐G‐F} n = 15; Cohort 2 Timepoint 2: App ^NL/NL n = 13, App ^{NL‐G‐F/NL‐G‐F} n = 13; Cohort 3: Wt littermates n = 9, App ^{NL‐G‐F/NL‐G‐F} n = 4; Cohort 4: Wt littermates n = 15, App ^{NL‐G‐F/NL‐G‐F} n = 9.

FIGURE 3

Male, but not female, App ^NL/NL controls are impaired at reversal learning compared with wildtype controls. Pairwise visual discrimination learning and reversal percent correct responses, number of correction trials, and perseveration index for the App ^NL/NL and wildtype controls from Cohorts 1 and 3 (younger cohorts) and Cohorts 2 (Timepoint 2) and 4 (older cohorts). (A) Genotypes, cohorts, ages, and images used during testing. There were no significant differences between the female groups at either age (B–I). However, despite acquiring the task at a similar rate (J), younger male wildtype mice were significantly impaired in percent correct responses during PVR than App ^NL/NL mice (K) but were no different in number of correction trials (L) or perseveration indices (M). Whereas older male mice acquired the task at a similar rate (N), App ^NL/NL mice made fewer correct responses (O), more correction trials (P), and had higher perseveration indices (Q) than wildtype control mice in both prior to (baseline) and following rule reversal. Note: Cohort 4 baseline trials were averaged here to conform with Cohort 2. Blue symbols and lines depict App ^NL/NL mice, yellow depict wildtype controls. Acquisition data depicted as mean ± SD. Baseline and reversal data depicted as mean ± SEM. *p < 0.05, **p < 0.01.

FIGURE 4

Female App ^{NL‐G‐F/NL‐G‐F} are unimpaired in extinction compared to App ^NL/NL controls. When compared with App ^NL/NL controls, App ^{NL‐G‐F/NL‐G‐F} female mice were unimpaired in extinction of a stimulus–response association. (A) Percent of responses made to the stimulus. (B) Number of touches to non‐active windows. (C) Number of touches to the response window made during the intertrial interval (ITI). (D–F) Percent responses during the first 10 trials (D), middle 10 trials (E), and last 10 trials (F) within each Session. Red symbols and line depict App ^{NL‐G‐F/NL‐G‐F} groups, blue depicts App ^NL/NL mice. App ^NL/NL n = 15, App ^{NL‐G‐F/NL‐G‐F} n = 14. Data depicted as mean ± SEM.

FIGURE 5

Male App ^{NL‐G‐F/NL‐G‐F} are unimpaired in extinction compared to App ^NL/NL controls. When compared with App ^NL/NL controls, App ^{NL‐G‐F/NL‐G‐F} male mice were unimpaired in extinction of a stimulus–response association. (A) Percent of responses made to the stimulus. (B) Number of touches to non‐active windows. (C) Number of touches to the response window made during the ITI. (D–F) Percent responses during the first 10 trials (D), middle 10 trials (E), and last 10 trials (F) within each Session. Red symbols and line depict App ^{NL‐G‐F/NL‐G‐F} groups, blue depict App ^NL/NL mice. App ^NL/NL n = 14, App ^{NL‐G‐F/NL‐G‐F} n = 15. Data depicted as mean ± SEM.