NEK4 modulates circadian fluctuations of emotional behaviors and synaptogenesis in male mice

Abstract

GWASs have linked the 3p21.1 locus, which is associated with the expression levels of NEK4, to bipolar disorder. Here, we use integrative analyses of GWAS statistics and eQTL annotations to establish that elevated NEK4 expression in the hippocampus is associated with an increased risk of bipolar disorder. To further study this association, we generate transgenic male mice that conditionally overexpress NEK4 in the pyramidal neurons of the adult forebrain, or use AAV to overexpress NEK4 in the dorsal hippocampus. Compared to the control mice, male mice of both strains exhibit a shift from a diurnal anxiety state to a nocturnal normal or anxiolytic-like state. Overexpression of NEK4 also affects the circadian fluctuations in dendritic spine morphology and synaptic structure. Furthermore, we show that treatment with lithium ameliorates the effects of NEK4 overexpression in male mice. We then perform phosphoproteomic analyses to demonstrate that the diurnal and nocturnal phosphoproteomic profiles of male control and NEK4 overexpressing mice are different. These results suggest that male mice with different NEK4 expression levels may recapitulate some of the core features observed in patients with bipolar disorder, indicating that interruption of the homeostatic dynamics of synapses may underlie the emotional swings in bipolar disorder.

Fig. 1. Genetic and molecular characterizations of 3p21.1 locus in PGC3 BD GWAS.

A Genetic associations of SNPs spanning 3p21.1 region with BD in European populations (41917 cases and 371549 controls). A physical map of the region is given and depicts known genes within the region, and three independent risk SNPs (rs7622851, rs2336147 and rs2276824) were marked. The linkage disequilibrium (LD) of the three SNPs were calculated in European individuals from 1000 Genomes Project, based on the r² algorithm implemented in SHEsis program. B Associations of rs7622851, rs2336147 and rs2276824 with mRNA expression of NEK4 in hippocampus (n = 371) from BrainSeq Phase 2 dataset. The BD risk-associated genotypes at each SNP were marked. In box plot B, the lines from top to bottom represent maximum, 3rd quartile, median, 1st quartile, and minimum, while the middle area represents the interquartile range. Source data are provided as a Source data file.

Fig. 2. Forebrain conditional transgenic mice (H11-LSL-NEK4^±; CaMKIIα-Cre) exhibited abnormal circadian fluctuations in emotional behaviors.

A Time scheme for mouse behavioral tests in NEK4 cTG mice. B Results of open field tests during the day and night of the NEK4 cTG mice compared to the control mice, including the total distance travelled, the distance spent in the center (day: P = 0.0393, df = 14, t = 2.272; night: P = 0.6931, df = 14, t = 0.403) and time spent in the center, and the time spent in the corner, were recorded and analyzed. C Results of the elevated plus maze during the day and night of the NEK4 cTG mice compared to the control mice, and their total distance traveled (day: P = 0.0008, df = 14, t = 4.280; night: P = 0.5681, df = 14, t = 0.585), distance spent in the open arms (day: P = 0.0019, df = 14, t = 3.801; night: P = 0.2647, df = 14, t = 1.162), time spent in the open arms (day: P = 0.0018, df = 14, t = 3.833; night: P = 0.0131, df = 14, t = 2.839) and time spent in the closed arms (day: P = 0.0006, df = 14, t = 4.363; night: P = 0.0238, df = 14, t = 2.535) were calculated. D The NEK4 cTG mice and the control mice underwent the forced swimming tests and monitored for 8 min, and the immobility time in the remaining 6 min was calculated for each mouse. N = 8 mice/group. All data were shown as mean ± SD, the two-sided student’s t test was used to perform the significance test. *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P < 0.0001. The shaded part represents the behavioral experiment data carried out in the dark environment at night (21:00-24:00). Source data are provided as a Source data file.

Fig. 3. AAV-mediated overexpression of NEK4 in the hippocampus led to aberrant circadian fluctuations in mood behaviors.

A Diagram showing the injection site and fluorescence expression in the dorsal hippocampus of mice (injection position [AP]: −2.06 mm; [ML]: ±1.50 mm; [DV]: −1.50 mm). B Time scheme for mouse behavioral tests in NEK4 OE mice. C Results of open field tests during the day and night of the NEK4 OE mice compared to the control mice. The total distance travelled, the time spent in the center (day: P = 0.0483, df = 29, t = 2.062; night: P = 0.4094, df = 29, t = 0.837), and the time spent in the corner (day: P = 0.0166, df = 29, t = 2.542; night: P = 0.4935, df = 29, t = 0.694) were recorded. Center area time during the day compared with the night in NEK4 OE group (P = 0.0029, df = 30, t = 3.241). N = 15 Control mice/group, n = 16 NEK4 OE mice/group. D Results of the elevated plus maze tests during the day and night of the NEK4 OE mice compared to the control mice. The total distance traveled, time spent in the open arms (day: P = 0.0299, df = 26, t = 2.297; night: P = 0. 0.0017, df = 28, t = 3.472), and distance traveled in the open arms (day: P = 0.0294, df = 26, t = 2.304; night: P = 0. 1056, df = 28, t = 1.672) were measured. Opened arm time and distance during the day compared with the night in NEK4 OE group (time: P < 0.0001, df = 27, t = 5.271; distance: P < 0.0001, df = 27, t = 4.090). Opened arm distance during the day compared with the night in control group (P = 0.0440, df = 27, t = 2.113). N = 14 mice/day group, n = 15 mice/night group. E Results of forced swimming tests during the day and night of the NEK4 OE mice compared to the control mice, the mice were placed in water for 8 min, and the lengths of their immobility time within the remaining 6 min were calculated (day: P = 0.909, df = 28, t = 0.116; night: P = 0.0337, df = 28, t = 2.234). N = 15 Control mice/group, n = 15 NEK4 OE mice/group. F The small animal metabolic and behavioral phenotype analysis system was used to monitor free activities of the mice for 2 days. N = 3 Control mice/group, n = 3 NEK4 OE mice/group. Data A-E were shown as mean ± SD, F was shown as mean ± SEM. The two-sided student’s t test was used to perform the significance test. *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P < 0.0001. The shaded part represents the behavioral experiment data carried out in the dark environment at night (21:00-24:00). Source data are provided as a Source data file.

Fig. 4. NEK4 overexpression disrupted circadian fluctuations in dendritic spine morphogenesis in the dorsal hippocampus CA1 area of mice.

A Sparsely labeled neurons in the dorsal hippocampus CA1 region are shown in the left panel. The injection site of NEK4 overexpression AAV was identified by red fluorescence, and gray fluorescence indicates neurons in the CA1 region labeled with sparsely labeled AAV. Scale bars represent 100 μm. The right panel shows the second or third branch in the apical dendrites of sparsely labeled neurons in the dorsal hippocampus CA1 region, and scale bars represent 5 μm. B The density of various types of dendritic spines in the apical dendrites of neurons from NEK4 OE mice and the control mice during day and night were calculated and compared. Dendritic spines during the day compared with the night in control group (thin: P = 0.0675, df = 31, t = 1.895; mushroom: P < 0.0001, df = 31, t = 5.009; two-sided student’s t test). Dendritic spines during the day compared with the night in NEK4 OE group (mushroom: P < 0.0001, df = 27, t = 6.251; two-sided student’s t test). Dendritic spines of the control group compared with NEK4 OE group (Day: thin: P < 0.0001, df = 99, t = 5.663; night: thin: P = 0.998, df = 75, t = 0.172. Day: mushroom: P = 0.0018, df = 99, t = 3.541; night: mushroom: P = 0.0008, df = 75, t = 3.815; 2-way ANOVA with Bonferroni’s multiple comparisons test). N = 19 neurons were analyzed in the daytime control group, n = 14 neurons were analyzed in the night control group, n = 16 neurons were analyzed in the daytime NEK4 OE group, n = 13 neurons were analyzed in the night NEK4 OE group. C The density of total dendritic spines from these mice were analyzed, the two-sided student’s t test was used to perform the significance test. All data were shown as mean ± SD, and P < 0.05 after corrections was defined as significant. *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P < 0.0001. Shadows represent data at night. Source data are provided as a Source data file.

Fig. 5. NEK4 overexpression disrupted circadian fluctuations in synaptogenesis in the dorsal hippocampus CA1 area of mice.

The number of synapses in NEK4 OE mice and the control mice during day and night were compared to assess the effects of NEK4 on synaptogenesis. A synapse was characterized based on the presence of the pre-protrusion (containing vesicles), synaptic cleft, and postsynaptic (PSD region) structures. The number of synapses in electron microscope images covering a same area was calculated and analyzed, and scale bars represent 2 μm. A Representative images of synaptic structures acquired by electron microscope. B The numbers of synapses in the CA1 region of dorsal hippocampus of the mice. Synapses during the day compared with the night in control group (P = 0.0022, df = 36, t = 3.301). Synapses during the day compared with the night in NEK4 OE group (P < 0.0001, df = 37, t = 5.187). Synapses of the control group compared with NEK4 OE group (Day: P < 0.0001, df = 36, t = 4.928; night: P = 0.0013, df = 37, t = 3.481). N = 19 images were analyzed in the daytime control group, n = 19 images were analyzed in the night control group, n = 19 images were analyzed in the daytime NEK4 OE group, n = 20 images were analyzed in the night NEK4 OE group. All data were shown as mean ± SD, the two-sided student’s t test was used to perform the significance test. *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P < 0.0001. Shadows represent data at night. Source data are provided as a Source data file.

Fig. 6. Lithium treatment ameliorated the circadian fluctuations in emotional behaviors in NEK4 OE mice.

A Mice treated with either vehicle or lithium were allowed to explore for 6 min in the open field, and the total distance traveled (day: control-NEK4 OE, P = 0.0354, df = 19, t = 2.265; night: control-NEK4 OE, P = 0.0233, df = 18, t = 2.479), distance traveled in the center area (day: control-NEK4 OE, P = 0.0324, df = 19, t = 2.308), time spent in the center area and time spent in the corner area (day: control-NEK4 OE, P = 0.0244, df = 19, t = 2.445) were calculated. B The mice were allowed to explore the elevated plus maze for 5 min, and the total distance traveled (day: control-NEK4 OE, P = 0.0046, df = 18, t = 3.237), distance traveled in the open arms (day: control-NEK4 OE, P = 0.1122, df = 18, t = 1.670; night: control-NEK4 OE, P = 0.0154, df = 18, t = 2.678), time traveled in the open arms (day: control-NEK4 OE, P = 0.0953, df = 18, t = 1.760; night: control-NEK4 OE, P = 0.0144, df = 18, t = 2.709), time spent in the closed arms (day: control-NEK4 OE, P = 0.0146, df = 18, t = 2.702; night: control-NEK4 OE, P = 0.0070, df = 18, t = 3.040) were calculated. N = 10 mice in the daytime of control group, n = 11 mice in the daytime of NEK4 OE group, n = 9 mice in the daytime of control group treated with lithium, n = 10 mice in the daytime of NEK4 OE group treated with lithium; n = 10 mice in the night of control group, n = 10 mice in the night of NEK4 OE group, n = 9 mice in the night of control group treated with lithium, n = 10 mice in the night of NEK4 OE group treated with lithium. All data were shown as mean ± SD, the two-sided student’s t test was used to perform the significance test. *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P < 0.0001. ^#P: Close to statistically significant. NEK4 OE: NEK4 overexpression. NEK4 OE-Li: Lithium treated NEK4 OE mice. Control-Li: Lithium treated control mice. The shaded part represents the behavioral experiment data carried out in the dark environment at night (21:00-24:00). Source data are provided as a Source data file.

Fig. 7. Lithium treatment ameliorated the circadian fluctuations in synaptic structures in NEK4 OE mice.

A Densities of various types of dendritic spines in apical dendrites neurons of dorsal hippocampus CA1 from NEK4 OE mice and the control mice treated with either vehicle or lithium. Dendritic spines of the control group compared with NEK4 OE group (day: thin: P = 0.0306, df = 138, q = 3.941; night: thin: P = 0.0417, df = 147, q = 3.774. Day: mushroom: P = 0.0085, df = 138, q = 4.559; night: mushroom: P < 0.0001, df = 147, q = 8.667; 2-way ANOVA with Tukey’s multiple comparisons test). N = 12 neurons in control-day; 13 neurons in NEK4 OE-day; 12 neurons in control-Li-day; 13 neurons in NEK4 OE-Li-day; 13 neurons in control-night; 13 neurons in NEK4 OE-night; 13 neurons in control-Li-night; 14 neurons in NEK4 OE-Li- night. B Effect of lithium treatment on synaptic formation in mice. synaptogenesis during the day compared with the night in control group (P = 0.0068, df = 56, t = 2.809; two-sided student’s t test); synaptogenesis of the control group compared with NEK4 OE group without lithium treatment (day: P = 0.0035, df = 55, t = 3.056; night: P = 0.0027, df = 56, t = 3.144; two-sided student’s t test). N = 28 images in control-day; 29 images in NEK4 OE-day; 31 images in control-Li-day; 28 images in NEK4 OE-Li-day; 30 images in control-night; 28 images in NEK4 OE-night; 28 images in control-Li-night; 29 images in NEK4 OE-Li-night. C Effect of lithium treatment on PSD thickness in mice. PSD thickness during the day compared with the night in control group (P = 0.0126, df = 43, t = 2.604; two-sided student’s t test); PSD thickness of the control group compared with NEK4 OE group without lithium treatment (day: P = 0.0173, df = 38, t = 2.489; night: P < 0.0001, df = 45, t = 4.940; two-sided student’s t test). N = 20 images in control-day; 20 images in NEK4 OE-day; 21 images in control-Li-day; 22 images in NEK4 OE-Li-day; 25 images in control-night; 22 images in NEK4 OE-night; 21 images in control-Li-night; 21 images in NEK4 OE-Li-night. All data were shown as mean ± SD, P < 0.05 after corrections was defined as significant. *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P < 0.0001. Shadows represent data at night. Source data are provided as a Source data file.

Fig. 8. Description of the diurnal proteins and nocturnal proteins in the control mice and NEK4 OE mice, as well as the Reactome gene set analysis and DisGeNET gene set analysis of nocturnal proteins.

In each box, proteins related to synapse and dendritic spines were shown; and the overlapped nocturnal proteins between the control mice and NEK4 OE mice were also listed.