Fig. 4

Differential phenotypes of netrin-G1 KO and netrin-G2 KO mice in learning and memory. a and b Object recognition test: a Latencies of netrin-G1-KO (3 mo-old) and netrin-G2 KO (3 mo-old) mice to approach novel and familiar objects were not significantly different. Mice were exposed to the open field on Day 1 (see Fig. 3). b Comparable to WT mice, both netrin-G1 KO and netrin-G2 KO mice spent a significantly greater percentage of time exploring the novel object than the familiar object during Day 3 (paired t-test: p < 0.01 for all groups). No genotype differences were detected. c and d Y-maze task: c Percent arm alternations in all groups was above the chance level (netrin-G1 KO, 13 mo-old; netrin-G2 KO, 13 mo-old). No genotype differences correlated with the percentage of spontaneous arm alterations. d Netrin-G2 KO mice exhibited a significant increase in the number of arm entries (Student’s t-test). e-h Spatial reference memory was examined using the Morris water maze task: e Netrin-G1 KO mice had a longer latency than WT mice to reach the hidden platform (6–7 mo-old; 2-way mixed ANOVA for genotype and day, ns for interaction; ***p < 0.001 post-hoc t-test). Netrin-G2 KO mice also had a longer latency (9–10 mo-old; 2-way mixed ANOVA for genotype, ns for interaction). f In the probe test, netrin-G1 KO and netrin-G2 KO mice spent significantly less time spent in the target quadrant (2-way ANOVA for genotype, ns for interaction, *P < 0.05, **P < 0.01, and *** < 0.001 post-hoc Bonferroni test). g Both netrin-G1 KO and netrin-G2 KO mice had significantly fewer crossings over the previous platform site (Student’s t-test). h The latency to reach the platform in the visible platform session was also significantly prolonged in netrin-G1 KO mice (2-way mixed ANOVA for genotype, ns for interaction). Netrin-G2 KO mice had a modestly shorter latency (2-way mixed ANOVA for genotype, ns for interaction). i Arm revisit errors were recorded in the radial maze task: netrin-G1 KO mice made a significantly greater number of errors in the early stage of training (3–4 mo-old; 2-way mixed ANOVA for genotype, ns for interaction). Netrin-G2 KO mice made significantly more revisit errors than WT mice in the all stages (4 mo-old; 2-way mixed ANOVA for genotype, ns for interaction) j: 5CSRTT was used to examine learning acquisition and spatial attention abilities: For netrin-G1 KO mice, the success rate was comparable between genotypes in session 1, and gradually and slightly declined in later sessions (7 mo-old; 2-way mixed ANOVA for genotype, ns for interaction). Netrin-G2 KO mice showed a significant delay in learning the task (7 mo-old; 2-way mixed ANOVA for genotype, ns for interaction). Data are presented as mean ± SEM