Design of anti-depressant phosphodiester amino acidic Keap1-Nrf2 protein-protein interaction inhibitors

Abstract

Inhibiting the protein-protein interaction (PPI) between Keap1 and Nrf2 is theoretically an effective and safe strategy for activation of Nrf2 pathway to treat major depressive disorder (MDD). In this study, through bioinformatic analysis of the brain tissues and peripheral blood of MDD patients and depressive mice, we confirmed the involvement of oxidative stress, inflammation, and the Keap1-Nrf2 pathway in depression. Subsequently, we developed a series of phosphodiester amino acidic diaminonaphthalene compounds as Keap1-Nrf2 PPI inhibitors for the first time. Screening using the LPS-stimulated SH-SY5Y and BV2 cell models identified compound 4-95 showing the best anti-oxidative stress and anti-inflammatory efficacy. The ability of 4-95 to penetrate the blood-brain-barrier was significantly enhanced. In a chronic unpredictable mild stress mouse model, treatment with 4-95 effectively ameliorated anxiety and depression behavior and restored serum neurotransmitter levels by promoting the Nrf2 nuclear translocation. Consequently, oxidative stress was reduced, and the expression of synaptic plasticity biomarkers, such as postsynaptic density protein 95 (PSD95) and synapsin 1 (SYN1) were significantly increased, suggesting the recovery of neuronal function. Collectively, our findings demonstrate that the Keap1-Nrf2 PPI inhibitor holds great promise as a preclinical candidate for the treatment of depression.

Keywords: Anti-depressant; Inhibitor; Keap1; Nrf2; Phosphodiester; Protein-protein interaction.

Fig. 1

Bioinformatic analysis of the brain tissues and peripheral blood of MDD patients. (A) Volcano plots of all the genes of the brain tissue from non-MDD subjects and MDD patients. (B) Volcano plots of all the genes in the peripheral blood from non-MDD subjects and MDD patients. (C) Heatmap of the Nrf2 targeted genes in the brain tissues from non-MDD subjects (n = 29) and MDD patients (n = 30). (D) Heatmap of the Nrf2 targeted genes in the peripheral blood from non-MDD (n = 4) and MDD patients with (n = 4). (E) GO enrichment analysis of up-regulated enriched DEGs in the brain tissues from non-MDD and MDD patients. (F) KEGG enrichment analysis of up-regulated enriched DEGs in the brain tissues from non-MDD subjects and MDD patients. (G) Up-regulated pathways in brain tissues of volunteers without MDD and patients with MDD by the KEGG-GSEA analysis. (H) The relationship between the down-regulated pathways obtained from the GSE101521 dataset analyzed by GSEA. The relationship between up-regulated pathways obtained from the GSE101521 dataset analyzed by GSEA, which were IL-17 signaling pathway (I), TNF signaling pathway (J), neuroactive ligand-receptor interaction (K), and NF-κB signaling pathway (L), .

Fig. 2

The biochemical assays of the compounds and the predicted binding modes of 4–95 with Keap1 Klech domain. (A–H) Determination of equilibrium dissociation constants K_D2 for each compound was performed using the fluorescence anisotropy assay. Data are mean ± SEM (n = 3). (I) The first binding pose of 4–95 with Keap1 Klech domain (The protein is from PDB 7XOT). (J) The second binding pose of 4–95 with Keap1 Klech domain. (K) The superposition of two possible binding poses of 4–95, NXPZ-2 (purple) and Keap1 Klech domain in a ribbon mode. Keap1 is shown in gray ribbon and the key residues of Keap1 interacting with compounds labeled and shown in the gray stick model. 4–95 is shown in the green or blue stick mode. Hydrogen bonds are depicted as light green dashed lines. The π-π stacking interactions are depicted as gray dashed lines. (L) Superposition of two binding poses of 4–95 and Keap1 in a surface mode.

Fig. 3

Effect of the compounds on the capabilities of anti-oxidation and anti-inflammation in LPS-induced SH-SY5Y cells. control: cells were cultured under normal condition without LPS nor compound. model: cells were treated with 100 ng/mL LPS. The compound number: cells were treated with 100 ng/mL LPS and 10 μM compound. (A) Immunofluorescence of SH-SY5Y cells under different culture conditions, stained with DCFH-DA (green). Scale bar = 200 μm. (B) The relative immunofluorescence intensity of ROS in SH-SY5Y cells under different culture conditions. Data are mean ± SEM (n = 6). The concentrations of SOD (C), GSH (D), and MDA (E) in the cell lysates with kits. The concentrations of IL-1β (F), TNF-α (G), and IL-10 (H) in the cell lysates with kits. Data are mean ± SEM (n = 4–6). ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001 vs control. #P < 0.05, ##P < 0.01, ###P < 0.001 vs model. && P < 0.01, &&& P < 0.001 vs 4–95. (I) The protein expression levels of total Nrf2, Keap1, HO-1, nuclear Nrf2, cytoplasm Nrf2, Lamin B1 and GAPDH in the cell lysates at different culture conditions. (J) The expression levels of Keap1 in SH-SY5Y cells (WT), SH-SY5Y cells transfected with lentiviral vector (NC), and SH-SY5Y cells with Keap1 KD. (K) The expression level of HO-1 in SH-SY5Y cells transfected with lentiviral vector (NC), and SH-SY5Y cells with Keap1 KD under different culture conditions. (L) The density ratio of Keap1 to β-actin. (M) The RNA levels of Keap1 in SH-SY5Y cells (WT), SH-SY5Y cells transfected with lentiviral vector (NC), and SH-SY5Y cells with Keap1 KD. Data are mean ± SEM (n = 3). ∗P < 0.05 vs WT. #P < 0.05 vs NC. (N) The density ratio of HO-1 to β-actin. Data are mean ± SEM (n = 3). ∗P < 0.05, ∗∗P < 0.01 vs NC. #P < 0.05 vs NC LPS.

Fig. 4

The effect of 4–95 on the depressive phenotype in the CUMS mouse model. first sentence of legend ‘were received’ appears twice - this should be corrected. control: mice were cultured under normal condition without modeling nor treatment. model: mice underwent the CUMS protocol. 4–95: mice were subjected to the CUMS model and administered 4–95. (A) Schematic of CUMS model and the schedule of treatment (20 mg/kg). (B) The total travel distance during OFT. (C) The central travel distance during OFT. (D) The grooming times during OFT. The trajectory in OFT (E) and elevated plus maze (EPM, F). (G) The percentage time spent in opened arm in EPM. (H) The percentage of open arm entries in EPM. (I) Latency to feeding in NSFT. (J) The food consumption in 30 min in NSFT. (K) The total fluid consumption in SPT. (L) The percentage of sucrose consumption in SPT. (M) The latency of the first immobility in TST. (N) The percentage of duration of total immobility inTST. The concentrations of corticosterone (O, R), NA (P, S), and 5-HT (Q, T) in the hippocampus and cortex with kits. Data are mean ± SEM (n = 8–17). ∗P < 0.05, ∗∗∗P < 0.001 vs control. #P < 0.05, ##P < 0.01, ###P < 0.001 vs model.

Fig. 5

Effects of 4–95 on the Keap1- Nrf2 pathway in the serum and brains of CUMS mice. control: mice were cultured under normal condition without modeling nor treatment. model: mice underwent the CUMS protocol. 4–95: mice were subjected to the CUMS model and administered 4–95. The concentrations of SOD (A), GSH (B), and MDA (C) in the mouse serum with kits. Data are mean ± SEM (n = 10). ∗P < 0.05, ∗∗∗P < 0.001 vs control. ###P < 0.001 vs model. The RNA levels of Nfe2l2 (D), Hmox-1 (E), Keap1 (F), and Nqo1 (G) in whole blood and PBMC. Data are mean ± SEM (n = 3). ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001 vs control. #P < 0.05, ##P < 0.01 vs model. (H) The protein expression levels of Nrf2, Keap1, HO-1, and NQO1 in the frontal cortex and hippocampus. (I) The Nrf2-ARE binding affinity determined by an EMSA in the hippocampus and cortex. (J) Immunofluorescence of brain slices of different groups stained with Nrf2 (red), Keap1 (green), and nuclear (DAPI, blue). Scale bar = 100/30 μm.

Fig. 6

The basic PK properties of 4–95, and the effect of 4–95 on the synaptic function marker proteins and BDNF pathway in the brains of CUMS mice. control: mice were cultured under normal condition without modeling nor treatment. model: mice underwent the CUMS protocol. 4–95: mice were subjected to the CUMS model and administered 4–95. (A) The ratio of brain compound concentration to blood compound concentration in male C57BL/6J mice. Data are mean ± SEM (n = 5/each time point). The PK parameters of 4–95 after intravenous (iv) (B) or intragastric (i.g.) administration (C) in SD rats. Data are mean ± SEM (n = 5). The expression levels of PSD95, and SYN1 in the frontal cortex (D, E, G) and hippocampus (D, F, H). The expression levels of BDNF, CREB, p-CREB, HDAC2, MeCP2, and mSin3A in the frontal cortex (D, I, K, M, O, Q) and hippocampus (D, J, L, N, P, R). Data are mean ± SEM (n = 3). ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001 vs control. #P < 0.05, ##P < 0.01 vs model.

Fig. 7

The effect and mechanism of 4–95 on the CUMS model.

Scheme 1

Synthetic route of POZL derivatives. Reagents and conditions: (a) Diethylphosphoric acetic acid, HATU, Pyridine, 85 °C, 3h, 87 %; (b) 2-bromoacetamide, K₂CO₃, KI, dry DMF, 60 °C, overnight, 61–79 %; (c) Different types of N-Boc amino acids, HATU, pyridine, 85 °C, 3 h, 45–67 %; (d) TFA, DCM, room temperature, 2 h, 91 %; (e) Diethylphosphoric acetic acid, HATU, Pyridine, 85 °C, 3h, 87 %.