Fig. 2

P021 treatment from embryonic day 8 through postnatal day 21 prevents the memory deficits at the age of 3–4 months in 3xTg-AD mice. Elevated plus maze (a–c), novel object recognition test (d, e), and Morris water maze (f–j) were carried out from day 90 to 120. Decrease in 3xTg-Vh mice but not in 3xTg-P021 mice of the number of arm entries (a) (two-way ANOVA, interaction F (1, 50) = 2.730 P = 0.1048; treatment F (1, 50) = 0.03329 P = 0.8560; genotype F (1, 50) = 33.15 P < 0.0001); b time spent in open arms (two-way ANOVA, interaction F (1, 52) = 0.3033 P = 0.5842; treatment F (1, 52) = 1.482 P = 0.2289; genotype F (1, 52) = 7.482 P = 0.0085); and c distance traveled in open arms (two-way ANOVA, interaction F (1, 51) = 4.500 P = 0.0688; treatment F (1, 51) = 0.6570 P = 0.4214; genotype F (1, 51) = 30.78 P < 0.0001). d No difference was observed in the percentage of time spent in exploring two identical objects during the 8-min sample phase (three-way ANOVA, genotypes × treatments × objects). e A 5-min test phase was carried out 20 min later, and the less-explored object during the sample phase was exchanged with a novel object. Discrimination index (time spent exploring the new object − time spent exploring the old object)/(time spent exploring both old and new objects) × 100% in the test phase was improved in 3xTg-P021 mice (two-way ANOVA, interaction, F (1, 42) = 17.46, P = 0.0001; treatment, F (1, 42) = 3.583, P = 0.0653; genotype, F (1, 42) = 0.04287, P = 0.8370; WT-Vh vs 3xTg-Vh P = 0.0483; WT-Vh vs WT-P021 P = 0.3703; 3xTg-Vh vs 3xTg-P021 P = 0.0007). f During the acquisition phase for 4 days, the escape latency to reach the hidden platform was increased in 3xTg-AD mice compared with WT controls, and treatment with P021 did not prevent it as determined by the three-way ANOVA but showed a clear tendency as determined by the two-tailed t test (three-way ANOVA, days, F (3, 192) = 91.77, P < 0.0001; genotype, F (1, 192) = 14.86, P = 0.0002; treatment, F (1, 192) = 2.596, P = 0.1088; two-tailed t tests, WT-Vh vs 3Tg-Vh, P < 0.01; 3xTg-Vh vs 3xTg-P021, P < 0.05). g In the probe trial, 3xTg-AD mice took more time in the first entrance into the target zone, and treatment with P021 decreased the latency significantly (two-way ANOVA, interaction F (1, 59) = 6.324 P = 0.0147; treatment F (1, 59) = 3.634 P = 0.0615; genotype F (1, 59) = 4.601 P = 0.0361; WT-Vh vs 3xTg-Vh P = 0.0095; WT-Vh vs WT-P021 P = 0.9737; 3xTg-Vh vs 3xTg-P021 P = 0.0132). h 3xTg-AD but not 3xTg-AD-P021 mice spent less time in the target quadrant than did WT mice (two-way ANOVA, interaction F (1, 56) = 0.8808 P = 0.3520; treatment F (1, 56) = 6.928 P = 0.0109; genotype F (1, 56) = 22.64 P < 0.0001; WT-Vh vs 3xTg-Vh P = 0.0007; WT-Vh vs WT-P021 P = 0.6163; 3xTg-Vh vs 3xTg-P021 P = 0.0745; two-tailed t tests, 3xTg-Vh vs 3xTg-P021 P = 0.0493). i 3xTg-AD mice and WT controls displayed similar swim speed. j Cued trial, examples of swim path to the cued platform; all animals found the cued platform. n = 10 for WT-Vh; n = 17 for WT-P021; n = 12 for 3xTg-Vh and n = 18 for 3xTg-P021. *P < 0.05, **P < 0.01, ***P < 0.001