Stress-induced reduction of Na+/H+ exchanger isoform 1 promotes maladaptation of neuroplasticity and exacerbates depressive behaviors

Abstract

Major depression disorder (MDD) is a neuropsychiatric disorder characterized by abnormal neuronal activity in specific brain regions. A factor that is crucial in maintaining normal neuronal functioning is intracellular pH (pHi) homeostasis. In this study, we show that chronic stress, which induces depression-like behaviors in animal models, down-regulates the expression of the hippocampal Na⁺/H⁺ exchanger isoform 1, NHE1, a major determinant of pHi in neurons. Knockdown of NHE1 in CA1 hippocampal pyramidal neurons leads to intracellular acidification, promotes dendritic spine loss, lowers excitatory synaptic transmission, and enhances the susceptibility to stress exposure in rats. Moreover, E3 ubiquitin ligase cullin4A may promote ubiquitination and degradation of NHE1 to induce these effects of an unbalanced pHi on synaptic processes. Electrophysiological data further suggest that the abnormal excitability of hippocampal neurons caused by maladaptation of neuroplasticity may be involved in the pathogenesis of this disease. These findings elucidate a mechanism for pHi homeostasis alteration as related to MDD.

Fig. 1.. CUMS induces depression-like behaviors in rats.

(A) SPT results showing that CUMS rats exhibited a decrease in percent of sucrose consumption. Control (Ctrl): 88.61 ± 1.70 versus CUMS: 54.59 ± 4.18. P < 0.001. (B and C) FST results showing that CUMS rats exhibited increased immobility times (Ctrl: 33.92 ± 5.17 versus CUMS: 126.80 ± 10.09; P < 0.001) and decreased swimming times (Ctrl: 123.80 ± 12.06 versus CUMS: 45.08 ± 5.39; P < 0.001). (D) Raw traces of rats in the OFT. (E) Total distance traveled in the OFT was not significantly different between control and CUMS rats. Ctrl: 7551.00 ± 597.10 versus CUMS: 6833.00 ± 553.40. P = 0.3869. (F) Time spent exploring the center area was decreased in CUMS rats. Ctrl: 7.24 ± 1.42 versus CUMS: 2.86 ± 0.52. P = 0.0080. (G) Raw traces of rats in the EPM. (H) Time spent in the open arms was decreased in CUMS rats. Ctrl: 7.69 ± 0.89 versus CUMS: 2.77 ± 0.36. P < 0.001. (I) Probability of entering open arms was decreased in CUMS rats. Ctrl: 29.24 ± 4.48 versus CUMS: 7.44 ± 1.36. P < 0.001. (J) Representative images of pHrodo Red staining in hippocampal neurons of rats. DAPI, 4′,6-diamidino-2-phenylindole. (K) Quantification analysis of the relative fluorescent intensity. Ctrl: 1.00 ± 0.06 versus CUMS: 1.33 ± 0.05. P = 0.0018. N = 6 per group. NS, not significant (P > 0.05); **P < 0.01 and ***P < 0.001 versus Ctrl by a two-tailed, unpaired Student’s t test.

Fig. 2.. CUMS causes maladaptation of synaptic plasticity in CA1 hippocampus of rats.

(A) Representative images of Golgi-stained hippocampal CA1 neurons. (B) Reconstructions of CA1 pyramidal neurons of rats. (C and D) Sholl analysis used to trace neuronal processes and showing the number of dendritic intersections. Ctrl: 94.50 ± 4.59 versus CUMS: 54.50 ± 3.19. P < 0.01, 15 neurons from N = 6 rats per group. (E) Golgi-stained dendrites. (F) Quantification of mushroom spines in the hippocampal neurons. Ctrl: 79.32 ± 2.21 versus CUMS: 57.06 ± 4.77. P = 0.0010, 45 dendritic segments from N = 8 to 9 rats per group. (G) Immunofluorescent staining showing VGLUT1⁺ (green) and PSD-95⁺ (red) colocalization in CA1 hippocampal neurons. (H) CUMS exposure decreased the number of synapses in neurons. Ctrl: 7120.00 ± 744.20 versus CUMS: 4841.00 ± 523.60. P = 0.0277, N = 5 per group. (I) Raw traces of mEPSC. (J) CUMS exposure decreased the amplitude of mEPSC. Ctrl: 15.27 ± 0.83 versus CUMS: 12.19 ± 0.96. P = 0.0280. (K) CUMS exposure decreased the frequency of mEPSC. Ctrl: 2.97 ± 0.20 versus CUMS: 1.85 ± 0.31. P = 0.0090. (L) Raw traces of sEPSC. (M) CUMS exposure decreased the amplitude of sEPSC. Ctrl: 31.52 ± 2.15 versus CUMS: 24.26 ± 1.06. P = 0.0061. (N) CUMS exposure decreased the frequency of sEPSC. Ctrl: 5.52 ± 0.43 versus CUMS: 2.77 ± 0.23. P < 0.01. (O) Raw traces showing individual voltage responses to a series of 700-ms current pulses from 0 to 140 pA with 20-pA steps. Red traces indicate the minimal current needed to induce APs. (P) Frequencies of induced APs at different current steps. N = 8 to 10 cells from six rats per group in electrophysiological recordings. *P < 0.05, **P < 0.01, and ***P < 0.001 versus Ctrl by a two-tailed, unpaired Student’s t test.

Fig. 3.. CA1 hippocampal neuronal NHE1 protein expression is decreased in CUMS rats.

(A) Schematic diagram of proteomic analysis. KEGG, Kyoto Encyclopedia of Genes and Genomes; GO, Gene Ontology. (B) Heatmap diagram of differentially expressed protein levels in proteomic analysis. (C) Volcano plots indicating differentially expressed proteins. FC, fold change. (D) The bar plot of GO analysis for NHE1 showed correlation to the regulation of pHi. CC, cellular component; MF, molecular function; BP, biological process. (E) Expression levels of NHE1 in hippocampal tissues was validated. Ctrl: 1.00 ± 0.03 versus CUMS: 0.70 ± 0.05. P = 0.0005, N = 6 per group. (F) Schematics of AAV vectors and bilateral injection sites in the hippocampus. Scale bar, 20 μm. eGFP, enhanced green fluorescent protein. L-ITR, left inverted terminal repeats; CaMKII, calcium/calmodulin dependent protein kinase II; CA, Cornu Ammonis. DG, dentate gyrus. (G) Validation of the efficiency by AAV-NHE1 short hairpin RNA (shRNA) injections. WT: 1.00 ± 0.08, AAV-eGFP: 0.84 ± 0.07 versus AAV-NHE1 shRNA: 0.55 ± 0.06. N = 6 per group. P = 0.0284, AAV-NHE1 shRNA versus AAV-eGFP. GAPDH, glyceraldehyde-3-phosphate dehydrogenase. (H) Knockdown of NHE1 caused rats more liable to decrease the percent of sucrose consumption. WT: 91.18 ± 1.35, AAV-eGFP: 84.62 ± 3.01, and AAV-NHE1 shRNA: 69.66 ± 4.41. N = 7 per group. (P = 0.0123, AAV-NHE1 shRNA versus AAV-eGFP). (I) Knockdown of NHE1 caused rats more liable to increase immobility times. WT: 29.13 ± 5.89, AAV-eGFP: 19.63 ± 4.31, and AAV-NHE1 shRNA: 115.40 ± 20.88. N = 8 per group. P < 0.001, AAV-NHE1 shRNA versus AAV-eGFP. (J) Knockdown of NHE1 caused rats more liable to decrease swimming times. WT: 109.30 ± 16.03, AAV-eGFP: 98.88 ± 10.93, and AAV-NHE1 shRNA: 53.38 ± 5.28. N = 8 per group. P = 0.0295, AAV-NHE1 shRNA versus AAV-eGFP. NS, P > 0.05; *P < 0.05; **P < 0.01; and ***P < 0.001 by a two-tailed, unpaired Student’s t test (F) and a one-way analysis of variance (ANOVA) with the Tukey’s post hoc correction (H to J). WT, wild type.

Fig. 4.. NHE1 knockdown results in decreased pHi and increased synaptic plasticity damage in CA1 hippocampal neurons.

(A) Representative images of pHrodo Red staining in CA1 hippocampal neurons. (B) Quantification analysis showing the pHi in CA1 pyramidal neurons. WT: 1.00 ± 0.06, AAV-eGFP: 1.14 ± 0.11, and AAV-NHE1 shRNA: 1.58 ± 0.08. N = 6 per group. P = 0.0084, AAV-NHE1 shRNA versus AAV-eGFP. (C) Golgi-stained CA1 hippocampal neurons. (D) Reconstructions of pyramidal neurons. (E and F) Quantification of dendritic intersections. N = 15 neurons from five rats per group. WT: 87.00 ± 5.36, AAV-eGFP: 88.83 ± 2.56, and AAV-NHE1 shRNA: 60.67 ± 3.05. P = 0.0003, AAV-NHE1 shRNA versus AAV-eGFP. (G) Representative images of dendrites in pyramidal neurons. (H) Quantification of mushroom spines in neurons. WT: 78.39 ± 3.68, AAV-eGFP: 77.21 ± 4.55, and AAV-NHE1 shRNA: 54.36 ± 4.92. N = 45 dendritic segments from six rats per group. P = 0.0062, AAV-NHE1 shRNA versus AAV-eGFP). (I) Raw traces of mEPSC in neurons. (J) NHE1 knockdown decreased amplitudes of mEPSC. WT: 14.78 ± 0.45, AAV-eGFP: 14.57 ± 0.31, and AAV-NHE1 shRNA: 9.22 ± 0.39. P < 0.001, AAV-NHE1 shRNA versus AAV-eGFP. (K) NHE1 knockdown decreased frequencies of mEPSC. WT: 4.23 ± 0.34, AAV-eGFP: 3.61 ± 0.31, and AAV-NHE1 shRNA: 2.02 ± 0.11. P = 0.0024, AAV-NHE1 shRNA versus AAV-eGFP. (L) Raw traces of sEPSC. (M) NHE1 knockdown decreased amplitudes of sEPSC. WT: 31.30 ± 2.62, AAV-eGFP: 29.98 ± 2.85, and AAV-NHE1 shRNA: 18.60 ± 1.84. P = 0.0142, AAV-NHE1 shRNA versus AAV-eGFP. (N) NHE1 knockdown decreased frequencies of sEPSC. WT: 4.95 ± 0.41, AAV-eGFP: 4.39 ± 0.40, and AAV-NHE1 shRNA: 2.14 ± 0.14. P = 0.0008, AAV-NHE1 shRNA versus AAV-eGFP. N = 5 to 6 cells from five rats per group in electrophysiological recordings. *P < 0.05, **P < 0.01, and ***P < 0.001.

Fig. 5.. CUMS activates CUL4A and ubiquitination degrades NHE1 protein within CA1 hippocampus.

(A) Levels of NHE1 mRNA were validated. Ctrl: 1.00 ± 0.05 versus CUMS: 0.91 ± 0.04. P = 0.1859, N = 12 per group. (B) GO function classification annotations indicated that the differentially expressed proteins appeared to be related with the posttranslational modification pathway. (C) Heatmap diagram of differentially expressed protein levels as related with posttranslational modification. (D) Chloroquine (CQ; 40 μM) failed to prevent CORT-induced (10 μM) decreases of NHE1 expression levels. Ctrl: 1.08 ± 0.06, CORT: 0.65 ± 0.08, and CORT + CQ: 0.73 ± 0.05. N = 6 per group. P = 0.0007, CORT versus Ctrl; P = 0.0046, CORT + CQ versus Ctrl; and P = 0.623, CORT + CQ versus CORT. (E) MG132 (20 μM) reversed the CORT-induced (10 μM) decreases of NHE1 expression levels. Ctrl: 0.92 ± 0.08, CORT: 0.45 ± 0.11, and CORT + MG132: 1.06 ± 0.07. N = 6 per group. P = 0.0039, CORT versus Ctrl; P = 0.5094, CORT + MG132 versus Ctrl; and P = 0.0004, CORT + MG132 versus CORT. (F) CUMS exposure increased CUL4A protein expression levels. Ctrl: 1.00 ± 0.06 and CUMS: 1.34 ± 0.05. P = 0.0019, N = 6 per group. (G) CUL4A and NHE1 were colocalized within the ER in primary cultured hippocampal neurons. **P < 0.01 and ***P < 0.001 by a two-tailed, unpaired Student’s t test (A and F) and a one-way ANOVA with the use of the Tukey’s post hoc correction (D and E).

Fig. 6.. CUL4A overexpression produces maladaptive plasticity in CA1 hippocampal pyramidal neurons.

(A) Experimental paradigm for viral construction and injection in Thy1-Cre mice. (B) Representative images of injection sites in CA1 hippocampus. Scale bar, 20 μm. (C) Expression of CUL4A (AAV-eGFP: 1.00 ± 0.05 versus AAV-CUL4A: 1.32 ± 0.08; P = 0.0074) and NHE1 (AAV-eGFP: 1.00 ± 0.09 versus AAV-CUL4A: 0.49 ± 0.10; P = 0.0033) in CA1 pyramidal neurons. (D) pHrodo Red staining in CA1 hippocampal neurons. (E) CUL4A overexpression decreased pHi in neurons. AAV-eGFP: 1.00 ± 0.05 versus AAV-CUL4A: 1.50 ± 0.06. P < 0.001, N = 6 per group. (F) Raw traces of electrophysiological recordings. (G) CUL4A overexpression decreased the amplitudes of sEPSC. AAV-eGFP: 1.00 ± 0.05 versus AAV-CUL4A: 1.50 ± 0.06. P = 0.0070, N = 5 to 6 cells from five rats per group. (H) CUL4A overexpression decreased the frequencies of sEPSC. AAV-eGFP: 7.52 ± 0.39 versus AAV-CUL4A: 4.12 ± 0.41. P = 0.0001, N = 5 to 6 cells from five rats per group. (I) Raw traces of evoked APs from injected currents (200 pA). (J) CUL4A overexpression decreased frequencies of APs. AAV-eGFP: 5.00 ± 0.84 versus AAV-CUL4A: 1.75 ± 0.48. P = 0.0166, N = 4 to 5 cells from four rats per group. (K) CUL4A overexpression decreased the percent of sucrose consumption. AAV-eGFP: 95.50 ± 0.32 versus AAV-CUL4A: 79.39 ± 3.14. P = 0.0009, N = 5 per group. (L) CUL4A overexpression increased immobility times in FST. AAV-eGFP: 26.17 ± 6.26 versus AAV-CUL4A: 109.20 ± 36.28. P = 0.0478, N = 6 per group. *P < 0.05, **P < 0.01, and ***P < 0.001 by a two-tailed, unpaired Student’s t test.

Fig. 7.. DREADD inhibition of CA1 pyramidal neurons produces depression-like behaviors in rats.

(A) AAV-DIO-hM4D(Gi) construct and injection site in CA1 hippocampal regions. Scale bar, 1 mm. IRES, internal ribosome entry site. (B) Representative voltage responses in CA1 pyramidal neurons to injection currents from 0 to 120 pA with 20-pA steps before and after bath application of CNO (5 μM). Red traces indicate the minimal current needed to induce APs. (C) CNO decreases spiking of hM4Di-expressing neurons. hM4Di: 52.00 ± 8.00 versus hM4Di + CNO: 156.00 ± 23.15. P = 0.0028. N = 5 cells from five mice per group. (D) Number of induced APs at different current steps. N = 5 cells from five mice per group. (E) Diagram of CNO (5 mg/kg) injection procedure. i.p., intraperitoneal. (F) FST results after saline or CNO injection in mice. mCherry + CNO: 35.00 ± 9.87, hM4Di + saline: 37.50 ± 8.29, and hM4Di + CNO: 111.90 ± 14.29. N = 6 per group, P = 0.0013. (G) TST results after saline or CNO injection in mice. mCherry + CNO: 37.57 ± 5.87, hM4Di + saline: 60.50 ± 2.90, and hM4Di + CNO: 120.00 ± 20.94. P = 0.0306, N = 4 per group. (H) Total distance in OFT after saline or CNO injection in mice. mCherry + CNO: 9538.00 ± 621.30, hM4Di + saline: 9316.00 ± 1506.00, and hM4Di + CNO: 9260.00 ± 1640.00. P = 0.9803, N = 6 per group. (I) Exploring times after saline or CNO injection in mice. mCherry + CNO: 16.20 ± 4.91, hM4Di + saline: 17.48 ± 2.94, and hM4Di + CNO: 9.11 ± 1.70. P = 0.0024, N = 6 per group. *P < 0.05, **P < 0.01, and ***P < 0.001 by a two-tailed, unpaired Student’s t test (F, G, I, and J).