General health
The general health check indicated no differences among all group [p > .050], except bald patch [c2(5) = 13.168, p = 0.022] (Table 1). A post-hoc analysis revealed that App-KI mice showed more bald patches than WT control [p = .001]. The neurological reflexes of all groups showed similar performances [p > .050] (Table 1).
Elevated plus maze
The state anxiety-like behavior of mice was assessed by measuring percentage of open arm entries out of total arm entries and time spent in the open arms. The mean percentage of open arm entries out of total arm entries in the elevated plus maze was higher in the App-KI group compared to WT group [Main effect of Gene: F (1, 53) = 17.017, p < .001, partial h2 = .243] (Fig. 2A). Percentage of open arm entries was not affected by general anesthesia exposure [Main effect of Anesthesia: F (2, 53) = 1.657, p = .200, partial h2 = .059], and there was no interaction between Gene and Anesthesia [Interaction: F (2, 53) = 0.609, p = .548, partial h2 = .022]. In the mean time spent in the open arm, the same tendency to percentage of open arm entries was observed [Main effect of Gene: F (1, 53) = 28.159, p < .001, partial h2 = .347; Main effect of Anesthesia: F (2, 53) = 0.493, p = .614, partial h2 = .018; Interaction: F (2, 53) = 0.135, p = .874, partial h2 = .005] (Fig. 2B). These results suggested aged App-KI mice exhibited excessive anxiolytic-like behavior and both desflurane and sevoflurane did not affect these performances.
Balance beam test
Motor coordination and balance were assessed by measuring the latency to traverse and the number of times the hind feet slipped off in each trial of balance beam test. The mean number of times the hind feet slipped off of each group was shown in Fig. 3A, 3B, and 3C. A three-way repeated ANOVA revealed that the mean number of slips was larger in App-KI group compare to WT group [Main effect of Gene: F (1, 51) = 32.507, p < .001, partial h2 = .389], and was decreased through the trials [Main effect of Trial: F (5, 255) = 14.328, p < .001, partial h2 = .219]. General anesthetics did not affect the number of slips [Main effect of Anesthesia: F (2, 51) = 0.095, p = .909, partial h2 = .004]. Although interactions between Gene and Anesthesia, Gene and Trial, and Anesthesia and Trial were not significant [p < .050], there was significant interaction between these three factors [F (10, 255) = 2.048, p = .041, partial h2 = .074]. A post-hoc analysis revealed that in all trials, anesthetic exposure did not affect the number of slips in App-KI group or WT group [p > .050]. Further, the number of slips in the 1st and 2nd trial of the air-treated App-KI group were larger than that after third trial [ps < .050], suggesting that motor learning occurred in App-KI group when not exposed to anesthesia. On the other hand, in both App-KI groups exposed to anesthesia, the number of slips did not decrease with each trial [ps > .050]. In all the trials, excluding the 2nd trial of sevoflurane exposure groups, anesthesia-exposed App-KI groups exhibit more slips than anesthesia-exposed WT groups [ps < .050], whereas air-treated App-KI group showed more slips than air-treated WT group only in 1st and 2nd trials [p < .010], suggesting that both desflurane and sevoflurane induced motor learning deficits in App-KI mice.
The mean latency to traverse was longer in App-KI group compare to WT group [Main effect of Gene: F (1, 51) = 8.271, p = .006, partial h2 = .140] (Fig. 3D, 3E, 3F). Similar to the number of slips, main effect of trial was significant [F (5, 255) = 46.216, p < .001, partial h2 = .475], and main effect of anesthesia was not significant [F (5, 51) = 0.625, p < .539, partial h2 = .024]. There was significant interaction between Gene and Trial [F (5, 255) = 3.945, p = .007, partial h2 = .072], but other interactions were non-significant [p > .050]. A post-hoc analysis revealed the latency to traverse in APP-KI group was longer in 1st and 2nd trials compare to 3rd, 4th, 5th and 6th trials, while the latency in WT group was longer in 1st trial compare to the other trials, suggesting retarded motor learning in APP-KI mice.
Tail suspension test
To assess the postural reflex and antidepressant-like activity, limb clasping score, total immobile duration, and latency to first immobile episode were measured in the tail suspension test. A Kruskal-Wallis test revealed significant differences (c2 (5) = 28.24, p < .001) in the limb clasping score (Fig. 4A). A post-hoc analysis revealed that App-KI group exhibit significant higher limb clasping score compared to WT group [U = 100, p < .001], whereas anesthesia did not affect limb clasping score [c2 (2) = 0.44, p = .804]. The mean immobile duration in the tail suspension test was longer in the APP-KI group compared to WT group [Main effect of Gene: F (1, 50) = 11.773, p < .001, partial h2 = .191] (Fig. 4B). Immobile duration was not affected by general anesthesia exposure [Main effect of Anesthesia: F (2, 50) = 0.508, p = .605, partial h2 = .020], and there was no interaction between Gene and Anesthesia [Interaction: F (2, 50) = 1.210, p = .307, partial h2 = .046].
In the mean latency to first immobile episode in the tail suspension test, the same tendency to immobile duration was observed [Main effect of Gene: F (1, 50) = 14.499, p < .001, partial h2 = .225; Main effect of Anesthesia: F (2, 50) = 0.620, p = .542, partial h2 = .024; Interaction: F (2, 50) = 0.160, p = .852, partial h2 = .006] (Fig. 4C). These results suggested aged App-KI mice exhibited deficits in posture reflex and earlier shift from active coping behavior to passive coping behavior, and both desflurane and sevoflurane did not affect these performances.