The type 3 adenylyl cyclase is required for novel object learning and extinction of contextual memory: role of cAMP signaling in primary cilia

Abstract

Although primary cilia are found on neurons throughout the brain, their physiological function remains elusive. Human ciliopathies are associated with cognition defects, and transgenic mice lacking proteins expressed in primary cilia exhibit defects in learning and memory. Recently, it was reported that mice lacking the G-protein-coupling receptor somatostatin receptor-3 (SSTR3), a protein expressed predominately in the primary cilia of neurons, have defective memory for novel object recognition and lower cAMP levels in the brain. Since SSTR3 is coupled to regulation of adenylyl cyclase, this suggests that adenylyl cyclase activity in primary cilia of CNS neurons may be critical for some forms of learning and memory. Because the type 3 adenylyl cyclase (AC3) is expressed in primary cilia of hippocampal neurons, we examined AC3(-/-) mice for several forms of learning and memory. Here, we report that AC3(-/-) mice show no short-term memory for novel objects and fail to exhibit extinction of contextual fear conditioning. They also show impaired learning and memory for temporally dissociative passive avoidance. Since AC3 is exclusively expressed in primary cilia, we conclude that cAMP signals generated within primary cilia contribute to some forms of learning and memory, including extinction of contextual fear conditioning.

Figure 1.

AC3 is highly expressed in the primary cilia of neurons in the CA1 region of the hippocampus. A, B, Representative images of staining for AC3 in primary cilia of neurons in the CA1 region of the hippocampus for AC3^+/+ mice (A) and AC3^−/− mice (B). Scale bar, 20 μm. n = 4 for each genotype. Green, AC3; blue, Hoechst nuclear staining.

Figure 2.

Hippocampal neurons in AC3^−/− mice express primary cilia. A–C, Representative images of AC3-immunoreactive cilia in the CA3 region of the hippocampus in AC3^+/+ mice labeled with antibodies against AC3 (green) (A), SSTR3 (red) (B), and colocalization of AC3 and SSTR3 (merge) (C). Scale bar, 20 μm. D–F, Representative images of immunoreactive cilia in the CA3 region of the hippocampus in AC3^−/− mice labeled with antibodies against AC3 (green) (D), SSTR3 (red) (E), and colocalization of AC3 and SSTR3 (merge) (F). Scale bar, 20 μm. n = 4 for each genotype. Blue, Hoechst nuclear staining.

Figure 3.

AC3^−/− mice show impaired TDPA learning and memory. A, AC3^−/− mice (n = 11) learned TDPA slower than AC3^+/+ mice (n = 14). The crossover latency for AC3^+/+ and AC3^−/− mice during training for TDPA differed significantly on day 2 (p = 0.005), day 3 (p = 0.028), and day 4 (p = 0.034) of training. **p < 0.01; *p < 0.05. Data are represented as means ± SEM. B, The percentage of AC3^−/− mice (n = 11) reaching the cutoff value (300 s) for crossover latency during the course of TDPA training was lower than AC3^+/+ mice (n = 14). C, Long-term memory for TDPA is impaired in AC3^−/− mice. Long-term memory was tested 21 d after animals reached the cutoff value (300 s) for crossover latency during TDPA training. AC3^−/− mice (n = 10) showed poorer retention of memory for TDPA training than AC3^+/+ mice (n = 14) (p = 0.001). **p < 0.01. Data are represented as means ± SEM. D, Short-term memory of AC3^−/− mice for TDPA training is impaired. The crossover latency (in seconds) of animals entering the dark chamber was monitored 10 min after a single training session for TDPA. There were significant difference between AC3^+/+ mice (n = 8) and AC3^−/− mice (n = 9) in crossover latency (in seconds) during the memory retention test (p = 0.014). *p < 0.05. Data are represented as means ± SEM.

Figure 4.

AC3^−/− mice fail to exhibit short-term memory for novel objects or extinction of contextual memory. A, AC3^+/+ mice (n = 6) but not AC3^−/− mice (n = 6) showed extinction of contextual memory during 7 d of extinction training [(day 5 (p = 0.015), day 6 (p = 0.023), and day 7 (p = 0.00004)]. **p < 0.01; *p < 0.05. Data are represented as means ± SEM. B, AC3^+/+ mice (n = 8) but not AC3^−/− mice (n = 7) showed short-term memory for a novel object. During training, both AC3^+/+ mice (n = 8) and AC3^−/− mice (n = 7) showed no preference for two identical objects (A1 and A2). However, 5 min after training, AC3^+/+ mice, but not AC3^−/− mice, showed a significant preference for the novel object (B) (p = 6.8 × 10⁻⁷). **p < 0.01. Data are represented as means ± SEM.