Ketamine improved LPS-induced depression-like behaviors
In the OFT, no significant difference in the total traveled distance within 5 min was observed in the four groups, indicating that LPS injection and ketamine administration did not affect the locomotor activity of mice (Fig. 2a; interaction: LPS × ketamine, F1,24 = 2.54, P > 0.05; LPS: F1,24 = 3.735, P > 0.05; ketamine: F1,24 = 3.663, P > 0.05; LPS is LPS + Sal vs Sal + Sal; ketamine is LPS + Ket vs LPS + Sal). LPS injection decreased the time spent in the center (interaction: LPS × ketamine, F1,24 = 4.40, P < 0.05; LPS: F1,24 = 8.869, P < 0.01; ketamine: F1,24 = 20.273, P < 0.01) and the number of entries into the center (interaction: LPS × ketamine; F1,24 = 18.253, P < 0.05; LPS: F1,24 = 5.116, P < 0.05; ketamine: F1,24 = 18.253, P < 0.01), both of these effects were rapidly reversed by ketamine administration (Fig. 2b). In the FST, LPS injection increased the immobility time, which was reversed by ketamine administration (Fig. 2c; interaction: LPS × ketamine, F1,24 = 15.13, P < 0.05; LPS: F1,24 = 13.414, P < 0.01; ketamine: F1,24 = 22.327, P < 0.01). In the NSFT, ketamine administration reversed LPS-induced increase in the feeding latency (interaction: LPS × ketamine, F1,24 = 21.226, P < 0.05; LPS: F1,24 = 30.552, P < 0.01; ketamine: F1,24 = 30.165, P < 0.01), and the total food consumption (interaction: LPS × ketamine, F1,24 = 1.126, P > 0.05; LPS: F1,24 = 4.504, P > 0.05; ketamine: F1,24 = 0.681, P > 0.05) in the four groups were not affected (Fig. 2d). Ketamine administration did not affect LPS-induced decrease in body weight change in the mice (Fig. 2e; interaction: LPS × ketamine, F1,32 = 0.871, P > 0.05; LPS: F1,32 = 0.959, P < 0.01; ketamine: F1,32 = 17.855, P > 0.05). LPS induced a poor coat state (indicated by a decreased score) in mice that was attenuated by ketamine administration (Fig. 2f; interaction: LPS × ketamine, F1,24 = 7.360, P > 0.05; LPS: F1,24 = 21.630, P < 0.01; ketamine: F1,24 = 19.293, P < 0.01). In summary, these behavioral results indicated that ketamine (10 mg/kg) administration can eliminate the depression and anxious-like behaviors caused by LPS injection without affecting the locomotor activity of mice.
Ketamine reversed LPS-induced extrasynaptic CaMKIIα activity in the hippocampus
In the extrasynaptic fractions of the hippocampus, ketamine administration reversed the elevated level of p-CaMKIIα induced by LPS injection (Fig. 3a; interaction: LPS × ketamine, F1,12 = 11.495, P < 0.05; LPS: F1,12 = 5.710, P < 0.05; ketamine: F1,12 = 5.604, P < 0.05). In the synaptic fractions of the hippocampus, no significant difference in the level of p-CaMKIIα was observed in the four groups (Fig. 3b; interaction: LPS × ketamine, F1,12 = 0.426, P > 0.05; LPS: F1,12 = 0.000, P > 0.05; ketamine: F1,12 = 0.015, P > 0.05).
Ketamine reversed LPS-mediated extrasynaptic GluN2B localization and phosphorylation in the hippocampus
In the extrasynaptic fractions, ketamine administration reversed LPS-induced increase in the level of GluN2B (Fig. 4a; interaction: LPS × ketamine, F1,12 = 3.267, P > 0.05; LPS: F1,12 = 7.045, P < 0.05; ketamine: F1,12 = 5.224, P < 0.05). In the synaptic fractions, no significant difference in GluN2B level was observed in the four groups (Fig. 4a; interaction: LPS × ketamine, F1,12 = 0.172, P > 0.05; LPS: F1,12 = 0.020, P > 0.05; ketamine: F1,12 = 0.775, P > 0.05). Next, to confirm this finding, dual antibody labeling of surface GluN2B (an antibody that specifically binds to the N-terminal of the GluN2B) and synapse-specific protein PSD95 was used to identify synaptic GluN2B localization. In the CA1, CA3, and DG of the hippocampus, no significant difference in GluN2B/PSD95 colocalization was observed in the four groups (Fig. 4e; CA1: interaction: LPS × ketamine, F1,20 = 0.109, P > 0.05; LPS: F1,20 = 0.981, P > 0.05; ketamine: F1,20 = 0.303, P > 0.05; CA3: interaction: LPS × ketamine, F1,20 = 2.381, P > 0.05; LPS: F1,20 = 0.000, P > 0.05; ketamine: F1,20 = 0.381, P > 0.05; DG: interaction: LPS × ketamine, F1,20 = 0.000, P > 0.05; LPS: F1,20 = 0.155, P > 0.05; ketamine: F1,20 = 0.843, P > 0.05). This result suggests that ketamine did not affect synaptic GluN2B localization. In the CA1 and DG of the hippocampus, ketamine administration reversed LPS-induced increase in GluN2B immunoreactivity (Fig. 4f; CA1: interaction: LPS × ketamine, F1,12 = 17.929, P < 0.05; LPS: F1,12 = 16.714, P < 0.01; ketamine: F1,12 = 14.044, P < 0.01; DG: interaction: LPS × ketamine, F1,12 = 33.842, P < 0.05; LPS: F1,12 = 52.299, P < 0.01; ketamine: F1,20 = 30.386, P < 0.01). Combined with the above results and given that ketamine administration did not affect synaptic GluN2B localization, this result suggests that the ketamine-mediated reduction in GluN2B mainly affects extrasynaptic fractions. In the extrasynaptic fractions, ketamine administration reversed LPS-induced increase in the level of p-GluN2B (Fig. 4b; LPS: H = 8.824, P < 0.05; ketamine: H = 8.824, P < 0.05). In the synaptic fractions, no significant difference in p-GluN2B level was observed in the four groups (Fig. 4b; interaction: LPS × ketamine, F1,12 = 0. 144, P > 0.05; LPS: F1,12 = 0.700, P > 0.05; ketamine: F1,12 = 0.096, P > 0.05). In the CA1, CA3, and DG of the hippocampus, no significant difference in p-GluN2B/PSD95 colocalization was observed in the four groups (Fig. 4g; CA1: interaction: LPS × ketamine, F1,20 = 0.773, P > 0.05; LPS: F1,20 = 0.057, P > 0.05; ketamine: F1,20 = 0.313, P > 0.05; CA3: interaction: LPS × ketamine, F1,20 = 2.381, P > 0.05; LPS: F1,20 = 0.000, P > 0.05; ketamine: F1,20 = 0.381, P > 0.05; DG: interaction: LPS × ketamine, F1,20 = 0.356, P > 0.05; LPS: F1,20 = 0.181, P > 0.05; ketamine: F1,20 = 0.007, P > 0.05). In the CA1 and DG of the hippocampus, ketamine administration reversed LPS-induced increase in p-GluN2B immunoreactivity (Fig. 4 h; CA1: interaction: LPS × ketamine, F1,12 = 21.120, P < 0.05; LPS: F1,12 = 12.588, P < 0.01; ketamine: F1,12 = 11.595, P < 0.01; DG: interaction: LPS × ketamine, F1,12 = 15.625, P < 0.05; LPS: F1,12 = 52.299, P < 0.01; ketamine: F1,20 = 10.632, P < 0.01). This result also suggests that the ketamine-induced reduction in p-GluN2B mainly affects extrasynaptic fractions. In summary, these results indicate that downregulation of GluN2B and p-GluN2B induced by ketamine is mainly derived from the extrasynaptic fractions of the hippocampus.
Ketamine reversed LPS-induced enhancement of the extrasynaptic interaction of p-CaMKIIα–GluN2B in the hippocampus
A GluN2B antibody was used to precipitate the NMDA receptors complex from the extrasynaptic fractions of the hippocampus. Immunoprecipitation assays revealed that p-CaMKIIα binds GluN2B receptors in the extrasynaptic fractions of the hippocampus (Fig. 5a). In addition, ketamine administration attenuated the enhancement of the interaction between p-CaMKIIα (interaction: LPS × ketamine, F1,12 = 8.372, P < 0.05; LPS: F1,12 = 8.207, P < 0.05; ketamine: F1,12 = 16.954, P < 0.01) and GluN2B (interaction: LPS × ketamine, F1,12 = 3.117, P > 0.05; LPS: F1,12 = 5.341, P < 0.05; ketamine: F1,12 = 9.571, P < 0.01) induced by LPS (Fig. 5b). These results showed that CaMKIIα was an important binding partner for GluN2B. These results were confirmed by CaMKIIα loss of function assays using siRNA. In cultured hippocampal neurons, after transfection with control siRNA and CaMKIIα siRNA, the efficiency of knocking down was evaluated by western blotting (Fig. 5e; t = 3.571, P < 0.05) and immunocytochemistry (Fig. 5g; t = 5.831, P < 0.0001). CaMKIIα siRNA reduced enzyme levels by about 30% (Fig. 5e, g). CaMKIIα siRNA decreased the level (Fig. 5d; t = 3.112, P < 0.05) and the number and fluorescence intensity of GluN2B puncta (Fig. 5i; U = 16, P < 0.0001; t = 2.516, P < 0.05) in the cultured hippocampal neurons. Taken together, these data suggest that the level and number of clusters of GluN2B were regulated by CaMKIIα signal.
Inhibition of CaMKIIα by KN93 prevented depression-like behaviors and reduced elevated extrasynaptic GluN2B localization and phosphorylation in the hippocampus induced by LPS
To further determine the relationship between the activation of CaMKIIα and GluN2B in vivo, an inhibitor of CaMKIIα, KN93, was used in the present study. In the NSFT, KN93 treatment prevented LPS-induced increase in feeding latency (interaction: LPS × KN93, F1,24 = 30.107, P < 0.01; LPS: F1,24 = 33.146, P < 0.01; KN93: F1,24 = 21.992, P < 0.01; LPS is LPS + DMSO versus Sal + DMSO; KN93 is LPS + KN93 versus LPS + DMSO), and the total food consumption in the four groups were not affected (Fig. 6a; interaction: LPS × KN93, F1,24 = 0.255, P > 0.05; LPS: F1,24 = 0.954, P > 0.05; KN93: F1,24 = 0.085, P > 0.05). Moreover, KN93 treatment prevented LPS-induced increase in immobility time in the FST (Fig. 6b; interaction: LPS × KN93, F1,24 = 22.909, P < 0.01; LPS: F1,24 = 13.584, P < 0.01; KN93: F1,24 = 5.873, P < 0.05). These behavioral results suggest that inhibition of CaMKIIα by KN93 results in an antidepressant phenotype. Furthermore, in the extrasynaptic fractions of the hippocampus, KN93 treatment prevented the increase in p-CaMKIIα levels induced by LPS (Fig. 6c; interaction: LPS × KN93, F1,22 = 5.312, P < 0.05; LPS: F1,22 = 4.039, P < 0.05; KN93: F1,22 = 15.795, P < 0.01). In the extrasynaptic fractions of the hippocampus, KN93 treatment prevented the increase in GluN2B level induced by LPS (Fig. 6d; interaction: LPS × KN93, F1,25 = 22.571, P < 0.05; LPS: F1,25 = 6.001, P < 0.05; KN93: F1,25 = 8.058, P < 0.01). In the synaptic fractions, no significant difference in GluN2B level was observed in the four groups (Fig. 6d; interaction: LPS × KN93, F1,12 = 0.002, P > 0.05; LPS: F1,12 = 3.494, P > 0.05; KN93: F1,12 = 0.428, P > 0.05). In the CA1, CA3, and DG of the hippocampus, no significant difference in GluN2B/PSD95 colocalization was observed in the four groups (Fig. 6 h; CA1: interaction: LPS × KN93, F1,20 = 3.375, P > 0.05; LPS: F1,20 = 1.042, P > 0.05; KN93: F1,20 = 0.042, P > 0.05; CA3: interaction: LPS × KN93, F1,20 = 0.632, P > 0.05; LPS: F1,20 = 0.632, P > 0.05; KN93: F1,20 = 0.040, P > 0.05; DG: interaction: LPS × KN93, F1,20 = 0.000, P > 0.05; LPS: F1,20 = 0.032, P > 0.05; KN93: F1,20 = 0.289, P > 0.05). In the CA1, CA3, and DG of the hippocampus, KN93 treatment reversed LPS-induced increased in GluN2B immunoreactivity (Fig. 6i; CA1: interaction: LPS × KN93, F1,12 = 21.889, P < 0.05; LPS: F1,12 = 12.144, P < 0.01; KN93: F1,20 = 15.884, P < 0.01; CA3: interaction: LPS × KN93, F1,12 = 8.833, P < 0.05; LPS: F1,12 = 13.404, P < 0.01; KN93: F1,20 = 16.046, P < 0.01; DG: interaction: LPS × KN93, F1,12 = 8.870, P < 0.05; LPS: F1,12 = 18.344, P < 0.01; KN93: F1,20 = 13.806, P < 0.01). In the extrasynaptic fractions, KN93 treatment reversed LPS-induced increased in p-GluN2B level (Fig. 6e; interaction: LPS × KN93, F1.12 = 9.784, P < 0.05; LPS: F1,12 = 5.871, P < 0.05; KN93: F1,12 = 18.922, P < 0.01). In the synaptic fractions, no significant difference in p-GluN2B level was observed in the four groups (Fig. 6e; interaction: LPS × KN93, F1,12 = 0.194, P > 0.05; LPS: F1,12 = 1.157, P > 0.05; KN93: F1,12 = 0.258, P > 0.05). In the CA1, CA3, and DG of the hippocampus, no significant difference in p-GluN2B/PSD95 colocalization was observed in the four groups (Fig. 6j; CA1: interaction: LPS × KN93, F1,20 = 0.169, P > 0.05; LPS: F1,20 = 0.169, P > 0.05; KN93: F1,20 = 0.061, P > 0.05; CA3: interaction: LPS × KN93, F1,20 = 2.609, P > 0.05; LPS: F1,20 = 0.290, P > 0.05; KN93: F1,20 = 0.290, P > 0.05; DG: interaction: LPS × KN93, F1,20 = 0.016, P > 0.05; LPS: F1,20 = 0.16, P > 0.05; KN93: F1,20 = 0.016, P > 0.05). In the CA1 and DG of the hippocampus, KN93 treatment reversed LPS-induced increased in p-GluN2B immunoreactivity (Fig. 6 k; CA1: interaction: LPS × KN93, F1,12 = 3.688, P < 0.05; LPS: F1,12 = 5.904, P < 0.05; KN93: F1,20 = 6.328, P < 0.05; DG: interaction: LPS × KN93, F1,12 = 24.451, P < 0.05; LPS: F1,12 = 29.510, P < 0.01; KN93: F1,20 = 12.616, P < 0.01). In summary, these results indicate that CaMKIIα activation has an effect on the localization and phosphorylation of GluN2B in the extrasynaptic fraction of the hippocampus and is related to the rapid antidepressant effect of ketamine.
Ketamine upregulated the expressions of p-CREB and BDNF and improved the synaptic dysfunction in the hippocampus
Ketamine administration blocked LPS-induced significantly decreased in p-CREB expression in the hippocampus (Fig. 7a; interaction: LPS × ketamine, F1,12 = 25.310, P < 0.05; LPS: F1,12 = 7.936, P < 0.05; ketamine: F1,12 = 6.672, P < 0.05). Ketamine administration blocked LPS-induced significantly decreased in BDNF expression in the hippocampus (Fig. 7b; interaction: LPS × ketamine, F1,12 = 22.039, P < 0.05; LPS: F1,12 = 6.525, P < 0.05; ketamine: F1,12 = 16.408, P < 0.01). Ketamine administration sufficiently prevented LPS-induced depression in SC-CA1 LTP of the hippocampus (interaction: LPS × ketamine, F1,8 = 10.706, P < 0.05; LPS: F1,8 = 7.476, P < 0.05; ketamine: F1,8 = 6.972, P < 0.05). Ketamine administration reversed the decrease in GluR1 level induced by LPS but did not affect GluR2 level (Fig. 7d, e; GluR1: interaction: LPS × ketamine, F1,12 = 21.165, P < 0.05; LPS: F1,12 = 4.478, P < 0.05; ketamine: F1,12 = 9.843, P < 0.01; GluR2: interaction: LPS × ketamine, F1,12 = 0.019, P > 0.05; LPS: F1,12 = 1.425, P > 0.05; ketamine: F1,12 = 0.010, P > 0.05).