High urea induces depression and LTP impairment through mTOR signalling suppression caused by carbamylation

Abstract

Background: Urea, the end product of protein metabolism, has been considered to have negligible toxicity for a long time. Our previous study showed a depression phenotype in urea transporter (UT) B knockout mice, which suggests that abnormal urea metabolism may cause depression. The purpose of this study was to determine if urea accumulation in brain is a key factor causing depression using clinical data and animal models.

Methods: A meta-analysis was used to identify the relationship between depression and chronic diseases. Functional Magnetic Resonance Imaging (fMRI) brain scans and common biochemical indexes were compared between the patients and healthy controls. We used behavioural tests, electrophysiology, and molecular profiling techniques to investigate the functional role and molecular basis in mouse models.

Findings: After performing a meta-analysis, we targeted the relevance between chronic kidney disease (CKD) and depression. In a CKD mouse model and a patient cohort, depression was induced by impairing the medial prefrontal cortex. The enlarged cohort suggested that urea was responsible for depression. In mice, urea was sufficient to induce depression, interrupt long-term potentiation (LTP) and cause loss of synapses in several models. The mTORC1-S6K pathway inhibition was necessary for the effect of urea. Lastly, we identified that the hydrolysate of urea, cyanate, was also involved in this pathophysiology.

Interpretation: These data indicate that urea accumulation in brain is an independent factor causing depression, bypassing the psychosocial stress. Urea or cyanate carbamylates mTOR to inhibit the mTORC1-S6K dependent dendritic protein synthesis, inducing impairment of synaptic plasticity in mPFC and depression-like behaviour. CKD patients may be able to attenuate depression only by strict management of blood urea.

Keywords: Carbamylation; Depression; LTP; Urea; mTOR.

Fig. 1

Urea accumulation impairs mPFC to induce depression in chronic kidney disease patients. a, Diagram of co-morbidity of depression secondary to CKD, COPD and CHF. Chi square test. b, Correlation between depressive symptoms and anxious symptoms. Red dotted line, cut-off value of Hamilton Depression Rating Scale. Red spots, each representing a CKD patient diagnosed as depression (n = 19). Black spots, each representing a CKD patient excluded from depression (n = 21). c, Serum urea level of participants (n = 15, 21, 19). d, Serum creatinine level of participants (n = 15, 21, 19). e, Screening of index on enlarged CKD patient's cohort. CKD patients without depression group was adjusted to 100%. Cre, creatinine; Gluc, glucose; Alb, albumin; Hb, hemoglobin; LDL, low-density lipoprotein; LP(a), lipoprotein(a); PTH, parathyroid hormone; TG, triglyceride; Chlo, chloride ions; K, potassium ions; Ca, calcium ions; P, phosphate ions. Specific numbers are in the Supplementary Table 5. f, Differences in mPFC resting-state functional connectivity between depressed CKD (CKD/Dep), non-depressed CKD (CKD/Nrm) and healthy control (Ctrl) groups (n = 20, 21, 19). mPFC ROI locations are in blue. Axial slice level indicated in the z-axis with 1 mm MNI space. Orange clusters represent regions of significantly higher connectivity in the first group, blue clusters significantly lower (the bar). Clusters obtained from t-tests of the seed-voxel correlation z-score maps for each pair of groups. Resulting t-score maps thresholded and clusterized at voxelwise p = 0.05, cluster α = 0.05.

Fig. 2

High urea causes depression in CKD mice. a–f Adenine-induced mouse model (n = 12). a, Serum urea level of mice. b, Serum creatinine level of mice. c, Urea level in mPFC. d, Sucrose preference test. e, Force swimming test. f, Coat score. One-way ANOVA test (a–f). Ctrl, control group. Ade, adenine-induced CKD mice group. Imi, imipramine administration. Ket, ketamine administration. Veh, vehicle (saline) administration.

Fig. 3

Uremic toxin urea induces depression-like behaviour, LTP impairment and synapses loss. a–i, j, A^+/+;B^+/+ mice vs A^+/+;B^−/− mice vs A^−/−;B^−/− mice. a, Serum urea level (n = 4). b, Urea level in mPFC (n = 4). c, Serum creatinine level (n = 4). d, Immobility in force swimming test (n = 12,12,11). e, 8-day sucrose preference test, w/w, water/water; s/s, sucrose/sucrose; s/w, sucrose/water (n = 11,12,11). f, Latency to food in novelty-suppressing feeding test (n = 6,6,7). g, Time in open arms in elevated plus maze (n = 8,8,6). h, Coat score (n = 9,12,8). i, Total grooming time in splash test (n = 10,11,10). j, Urea (20 mM) suppressed LTP in mPFC. Left, representative traces and LTP time course. Calibration bars are 50 pA, 50 ms; right, summary graphs of LTP magnitude. k, Expression of PSD95, Synapsin-1 and MAP2 in vivo by Western blot. l, Dose-dependent effect of urea on expression of PSD95, Synapsin-1, and MAP2 in vitro by Western blot. One-way ANOVA test (a–i), two-way ANOVA test (j).

Fig. 4

mTORC1-S6K pathway is implicated in the effect of urea. a, mTORC1-S6K pathway in mPFC of adenine-induced CKD mice by Western blot. b, mTORC1-S6K pathway in mPFC of A^+/+;B^+/+ mice, A^+/+;B^−/− mice and A^−/−;B^−/− mice by Western blot . c, Dose-dependent effect of urea on mTORC1-S6K pathway in vitro by Western blot. d, Schematic of the AAV vector encoding constitutive active S6K (S6K^ac), vehicle, and the injection site into the mPFC (white bar, 20 μm). e, Phosphorylation level of S6 and expression of PSD95, Synapsin-1 and MAP2 on mPFC of A^+/+;B^−/− mice injected of S6K^ac or vehicle by Western blot. f, Urea (20 mM) could not suppress LTP on slices expressing S6K^ac. Left, representative traces and LTP time course. Calibration bars are 50 pA, 50 ms; right, summary graphs of LTP magnitude. g–k, Behaviour tests of A^+/+;B^−/− mice injected of S6K^ac or vehicle virus in mPFC. g, Sucrose preference test (n = 6). h, Time in open arms in elevated plus maze (n = 6). i, Immobility in force swimming test (n = 4,5). j, Latency to food in novelty-suppressing feeding test (n = 6). k, Coat score (n = 8). l–p, Behaviour tests of A^+/+;B^−/− mice administrated with ketamine or vehicle. l, Sucrose preference test. m, Time in open arms in elevated plus-maze. n, Immobility in force swimming test. o, Latency to food in novelty-suppressing feeding test. p, Coat score. (l–p, n = 5.) Student's t-test (g–p), two-way ANOVA test (f).

Fig. 5

Mechanism and target of urea. a-g, Effect of cyanate on mice. a, Serum urea level in cyanate-feeding mice (n = 11,12). b, mPFC urea level in cyanate-feeding mice. c, Sucrose preference test. d, Immobility in force swimming test. e, Time in open arms in elevated plus-maze. f, Immobility in tail suspension test. g, Coat score. (b–g, n = 12.) h, Cyanate (2 mM) suppressed LTP in mPFC. Left, representative traces and LTP time course. Calibration bars are 50 pA, 50 ms; right, summary graphs of LTP magnitude. i, Changes of mTORC1-S6K pathway and markers of synapses in mPFC of cyanate-feeding mice by Western blot. j, Dose-dependent effect of cyanate on mTORC1-S6K pathway and markers of synapses in vitro by Western blot. k–m, Target of urea/cyanate, carbamylation pull-down on proteins distracted from mPFC. k, Adenine-induced CKD mice. l, A^+/+;B^+/+ mice, A^+/+;B^−/− mice. m, Cyanate-feeding mice. Student's t-test (a–g), two-way ANOVA test (h).