Depletion of primary cilia from mature dentate granule cells impairs hippocampus-dependent contextual memory

Abstract

The primary cilium, a sensory organelle, regulates cell proliferation and neuronal development of dentate granule cells in the hippocampus. However, its role in the function of mature dentate granule cells remains unknown. Here we specifically depleted and disrupted ciliary proteins IFT20 and Kif3A (respectively) in mature dentate granule cells and investigated hippocampus-dependent contextual memory and long-term plasticity at mossy fiber synapses. We found that depletion of IFT20 in these cells significantly impaired context-dependent fear-related memory. Furthermore, we tested synaptic plasticity of mossy fiber synapses in area CA3 and found increased long-term potentiation upon depletion of IFT20 or disruption of Kif3A. Our findings suggest a role of primary cilia in the memory function of mature dentate granule cells, which may result from abnormal mossy fiber synaptic plasticity. A direct link between the primary cilia of mature dentate granule cells and behavior will require further investigation using independent approaches to manipulate primary cilia.

Figure 1. Depletion of primary cilia from mature dentate granule cells.

(a) Schematic diagram of the AAV injection in the dentate gyrus. (b) Representative confocal images of the hippocampus of IFT20 fl/fl mice transduced with AAV-CaMKII-eGFP or AAV-CaMKII-eGFP-Cre virus at 28 days post AAV labeling. Scale bar: 200 μm. (c) Representative confocal images of the granule cell layer of AAV-CaMKII-eGFP or AAV-CaMKII-eGFP-Cre 28 days post injection. Immunostaining with anti DCX (red) and ACIII (white) reveals that primary cilia were selectively ablated in mature DGCs by AAV-CaMKII-eGFP-Cre transduction, whereas no ciliary disruption was found by AAV-CaMKII-eGFP transduction. Scale bars: 30 μm (d) AAV-CaMKII-eGFP-Cre transduced cells exhibited a significantly depressed rate of ACIII expression versus AAV-CaMKII-eGFP controls. (Cre: 11.7 ± 2.1% ACIII+, eGFP: 82.6 ± 7.4% ACIII+; two-tailed un-paired t-test p = 8.5 × 10⁻⁷). (e) Cells in the granule cell layer that did not take up virus expressed primary cilia at comparable rates between eGFP control and Cre groups (CTRL: 82.6 ± 7.0%; Cre: 77.3 ± 7.2%; two-tailed unpaired t-test p = 0.634; n = 3,3). (f) Cre transduction resulted in significant gross depletion of primary cilia. (Averaged number of ACIII positive cells per field in CTRL and IFT20(−/−)^mDGCs: 47.23 ± 2.78 for CTRL and 1.8 ± 0.28 for IFT20(−/−)^mDGCs; p = 5.985 × 10⁻¹⁵; n = 4,4). (g) Expression of ACIII in DCX+ cells did not differ significantly between eGFP-Cre and eGFP controls (CTRL: 9.5 ± 3.6% co-localization; IFT20(−/−)^mDGCs : 12.8 ± 2.9% co-localization; two-tailed unpaired t-test p = 0.50). Fields size: 320 μm × 320 μm × 36 μm. ***p < 0.001; ****p < 0.0001; n is the number of animals.

Figure 2. Primary cilia ablation from mature dentate granule cells impairs contextual fear memory.

(a) Experimental design for CFC test. (b) IFT20(−/−)^mDGCs mice showed a decreased freezing response on the CFC test (repeated measures ANOVA main effect of condition: F(1,16) = 6.979, p = 0.018; n = 17,19). (c) IFT20(−/−)^mDGCs and CTRL mice did not differ significantly on the tone-cued fear memory test (two-way ANOVA main effect of condition: F(2,69) = 2.444; p = 0.094; n = 17,19). *p < 0.05; n is the number of animals.

Figure 3. Ablation of primary cilia from mature dentate granule cells impairs pre-exposure contextual fear memory.

(a) Experimental design for PECFC test. (b) IFT20(−/−)^mDGCs showed a reduced freezing response as compare to CTRL when re-exposed to a fear context (repeated measures ANOVA interaction between condition and time F(1, 25) = 4.747, p < 0.0001, followed by LSD comparison at 60–120 sec: p = 4.41 × 10⁻⁵; n = 22,23). ****p < 0.0001; n is the number of animals.

Figure 4. Removal of primary cilia from mature dentate granule cells increases synaptic plasticity in the mossy fiber pathway.

(a) A schematic representation of the tri-synaptic circuit and the position of stimulating and recording electrodes. (b) Top, representative traces of both CTRL and IFT20(−/−)^mDGCs recorded before and after HFS. Bottom, IFT20(−/−)^mDGCs show increased LTP compare to CTRL. (c) Average potentiation of fEPSPs slope during 50–60 minute showing enhanced LTP in IFT20(−/−)^mDGCs (two-tailed unpaired t-test p = 0.048; n = 7,8). (d) Top, representative traces of both WT and dnKif3A recorded before and after HFS. Bottom, dnKif3A show increased LTP compare to WT. (e) Average potentiation of fEPSPs slope during 50–60 minute showing LTP enhancement in dnKif3A (two-tailed unpaired t-test p = 0.048; n = 5, 4). *p < 0.05; n is the number of animals.