Somatostatin signaling in neuronal cilia is critical for object recognition memory

Abstract

Most neurons possess a single, nonmotile cilium that projects out from the cell surface. These microtubule-based organelles are important in brain development and neurogenesis; however, their function in mature neurons is unknown. Cilia express a complement of proteins distinct from other neuronal compartments, one of which is the somatostatin receptor subtype SST(3). We show here that SST(3) is critical for object recognition memory in mice. sst3 knock-out mice are severely impaired in discriminating novel objects, whereas they retain normal memory for object location. Further, systemic injection of an SST(3) antagonist (ACQ090) disrupts recall of familiar objects in wild-type mice. To examine mechanisms of SST(3), we tested synaptic plasticity in CA1 hippocampus. Electrically evoked long-term potentiation (LTP) was normal in sst3 knock-out mice, while adenylyl cyclase/cAMP-mediated LTP was impaired. The SST(3) antagonist also disrupted cAMP-mediated LTP. Basal cAMP levels in hippocampal lysate were reduced in sst3 knock-out mice compared with wild-type mice, while the forskolin-induced increase in cAMP levels was normal. The SST(3) antagonist inhibited forskolin-stimulated cAMP increases, whereas the SST(3) agonist L-796,778 increased basal cAMP levels in hippocampal slices but not hippocampal lysate. Our results show that somatostatin signaling in neuronal cilia is critical for recognition memory and suggest that the cAMP pathway is a conserved signaling motif in cilia. Neuronal cilia therefore represent a novel nonsynaptic compartment crucial for signaling involved in a specific form of synaptic plasticity and in novelty detection.

Figure 1.

Validation of SST₃-specific antibody confirms cilia localization of the receptor. A–F, Immunostaining in hippocampus of wild-type mice (A, C, E) and sst3 knock-out mice (B, D, F). A, B, Hippocampus; C, D, dentate gyrus; E, F, CA1. Scale bars: A (for A, B), 200 μm; C (for C, D), 100 μm; E (for E, F), 10 μm.

Figure 2.

Normal appearance of cilia in CA1 of sst3 knock-out mice. Immunostaining with anti-ACIII (red) antibody reveals no gross disruption of cilia structure or number in sst3 knock-out (B) versus wild-type (A) mice. DAPI (nuclear) staining is in blue.

Figure 3.

Impairment of ORM in sst3 knock-out mice. A, sst3 knock-out mice show no impairment of ORM with a 5 min retention interval between the sample phase and the test phase. Dashed lines in this and subsequent figures indicate equal exploration of both objects. B, With a 1 h delay between the sample phase and the test phase, sst3 knock-out mice, in contrast to wild-type mice, do not show a preference for the novel object. *Significant preference (one-sample t test); ^¥significant difference from wild type (unpaired t test).

Figure 4.

Overtraining does not restore deficit in ORM in sst3 knock-out mice. The training protocol consisted of three 5 min trials with a 10 min intertrial interval. After a 1 h delay, mice were tested for their preference for a novel object. *Significant preference (one-sample t test); ^¥significant difference from wild type (unpaired t test).

Figure 5.

sst2, sst4, and sst2/sst4 double knock-outs show normal ORM. A, When tested with a 5 min retention interval, sst2 and sst4 knock-out mice are equal to wild-type mice in the ability to detect a novel object. B, sst2, sst4, and sst2/sst4 knock-out mice show normal ability to discriminate a novel object after a 1 h retention interval. *Significant preference index (one-sample t test).

Figure 6.

Injection of SST₃ antagonist disrupts ORM after a 1 h retention interval in wild-type mice. A, SST₃ antagonist ACQ090 injected intraperitoneally 30 min before sample phase blocks novel object discrimination at a dose of 50 μg/kg. B, Intraperitoneal injection of ACQ090 30 min before test phase blocks recall of object recognition memories at doses of 10, 20, and 50 μg/kg. *Significant preference (one-sample t test); ^¥significant difference from vehicle injections (unpaired t test).

Figure 7.

Memory for object location is normal in sst3 knock-out mice. When location of an object was moved from its position during training, sst3 knock-out mice showed equal preference for the displaced object as wild-type mice. A, 1 h retention interval. B, 2 h retention interval.

Figure 8.

CA1 synaptic responses and electrically stimulated LTP in sst3 knock-out mice. A, A plot of stimulus strength versus synaptic response shows basic excitatory neurotransmission in CA1 is normal in sst3 knock-outs. Representative voltage traces show response at three different stimulus intensities in wild-type and sst3 knock-out mice. B, Paired-pulse facilitation at 15–200 ms interstimulus intervals is normal in sst3 knock-out mice. Representative voltage traces at 50 ms interstimulus intervals in wild-type and sst3 knock-out mice. C, CA1 LTP generated by stimulation of the Schaeffer collaterals with two HFTs is normal in sst3 knock-outs. D, Application of SST₃ agonist L-796,778 (1 μm) increases LTP induced by two HFTs in wild-type mice. Red bar shows time of application (0–16 min). Insets show representative fEPSPs for each condition: black traces are baseline, blue traces are 60 min following the trains. In this and subsequent figures, stimulus artifacts are digitally truncated. Calibration: 0.5 mV, 10 ms.

Figure 9.

SST₃ is critical for mediation of LTP induced by activation of AC. A, Forskolin-stimulated LTP is impaired in sst3 knock-out mice. Forskolin/IBMX stimulates some acute synaptic potentiation in sst3 knock-outs, but little sustained potentiation is maintained after washout. Blue bar indicates time of forskolin/IBMX application. B, Superfusion of SST₃ antagonist ACQ090 (0.5 μm; green bar) before and during forskolin application (blue bar) limits generation of LTP. When ACQ090 is applied following the forskolin application (orange bar), no LTP deficit is observed. Insets show representative fEPSPs for each condition, black traces are baseline, gray traces are 60 min following the washout of the forskolin mixture. Calibration: 0.5 mV, 10 ms.

Figure 10.

Effects of SST₃ agonist and antagonist in hippocampal lysate and slices. Forskolin (Fsk) stimulation of cAMP is inhibited by 1 μm ACQ090 in hippocampal slices but not hippocampal lysate. The SST₃ agonist L-796,778 (L-778) increases basal cAMP levels in hippocampal slices but not hippocampal lysate. *Significant difference from control cAMP levels (basal or forskolin-stimulated in the absence of SST₃ ligand; dashed line).