Dynorphin/KOR signaling in the PL is required for CSDS-induced depressive-like behaviors
CSDS is a widely used preclinical model for major depressive disorder (MDD), inducing depressive-like syndromes, such as social avoidance, behavioral despair and anhedonia. We observed that mice subjected to a 10-day CSDS regimen exhibited a significant depression phenotype. This was characterized by a decreased social interaction ratio in the social interaction test (SIT), increased immobility time in the tail suspension test (TST) and forced swimming test (FST), and decreased sucrose consumption in the sucrose preference test (SPT), while their locomotor activity remained unaffected (Fig. 1A-G). We previously identified that the claustrum-prelimbic cortex (PL) circuit plays a crucial role in modulating depressive-like behaviors induced by CSDS, accompanied by an increase in the endogenous opioid peptide dynorphin A levels within the PL [32]. To investigate the detailed molecular mechanisms involved in dynorphin/KOR activation and their contribution to depressive-like behaviors, we first confirmed the changes in the expression levels of dynorphin A in the PL. Our observations revealed elevated levels of dynorphin A in the PL following CSDS (Fig. 1H), but not in other mPFC subregions such as the anterior cingulate cortex (ACC) and infralimbic cortex (IL) (Figure S1).
To test whether subsequent KOR activation is responsible for CSDS-induced depressive-like behaviors, we administered the KOR antagonist norBNI into PL. As shown in Fig. 1I-O, norBNI significantly inhibited CSDS-induced depressive-like behaviors by increasing the social interaction ratio in the SIT, reducing immobility time in the TST and FST, and enhancing sucrose consumption in SPT, without affecting the locomotor activity. Further, using a genetic approach, we selectively knocked down KOR expression in the PL via stereotaxic injection of AAV2/9-CAG-mCherry-T2A-Cre (AAV-mCherry-Cre) or AAV2/9-CAG-mCherry-2A (AAV-mCherry) into the PL of KOR mice (Fig. 1P, Q). Knockdown efficacy was confirmed by immunoprecipitation (Fig. 1R, S). As shown in Fig. 1T-X, mice with KORs knockdown in the PL alleviated CSDS-induced depressive-like behaviors, which were similar to the effects observed with norBNI treatment. Additionally, global KOR knockout (KOR) mice did not exhibit significant depressive-like behaviors after CSDS compared to wild-type counterparts (Figure S2A-F). To ascertain whether PL-specific KORs are sufficient for stress-induced depressive-like behaviors, we microinjected AAV-EGFP-KOR or AAV-EGFP into the PL of KOR mice (Figure S2G, H). Viral expression of KOR was confirmed via immunoprecipitation (Figure S2I). As shown in Figure S2J, restoration of KOR expression in the PL induced significant social defeat stress-induced social avoidance. These findings indicate that dynorphin/KOR signaling in the PL is both necessary and sufficient for the development of CSDS-induced depressive-like behaviors.
P38 MAPK is a significant downstream effector of KOR activation and mediates the aversive properties of stress [33, 35, 36]. We next investigated the involvement of p38 MAPK in KOR-mediated depressive-like behaviors induced by CSDS. We detected the expression level of p38 MAPK phosphorylation. As shown in Fig. 2A, C, we found that CSDS significantly increased p38 MAPK phosphorylation in the PL.
Glutamatergic dysfunction was critically implicated in depressive disorders. The glutamate homeostasis is regulated by a variety of Na-dependent and independent transporters, including GLT-1, EAAT3, and xCT. We thus detected the levels of GLT-1, EAAT3, and xCT in the PL after CSDS. As shown in Fig. 2B, D, social defeat stress led to a marked downregulation of xCT expression in the PL, without affecting other transporters such as GLT-1 and EAAT3 (Figure S3). To determine the KOR dependence of these effects, we administered the KOR antagonist norBNI into the PL and found that norBNI reversed stress-induced p38 MAPK activation and restored xCT levels (Fig. 2E, F). Similarly, knockdown of KORs in the PL also significantly decreased p38 MAPK phosphorylation and increased xCT expression following exposure to CSDS (Fig. 2G, H).
Further, the role of p38 MAPK activation and xCT downregulation in CSDS-induced depressive-like behaviors were investigated. We microinjected p38 MAPK inhibitor Adezmapimod (SB203580) into the PL and found that SB203580 significantly attenuated CSDS-induced social avoidance in SIT, increment of immobility in FST, and decrement of sucrose consumption in SPT, without affecting locomotor activity (Fig. 2I-N). SB203580 also reversed stress-mediated increases in p38 phosphorylation and reductions in xCT expression (Figure S4). We explored xCT modulation using N-acetylcysteine (NAC), a cysteine prodrug and glutathione (GSH) precursor [12, 37]. As shown in Fig. 2O-T, intra-PL microinjection of NAC significantly inhibited CSDS-induced social avoidance in SIT, an increase of immobility in FST, and a decrease of sucrose consumption in SPT, with no impact on locomotor activity. In addition to being an xCT activator, NAC is widely used as a pharmacological antioxidant and cytoprotectant. To exclude its antioxidant potential in its antidepressant effects, we detect the reactive oxygen species (ROS) level after stress. It was found that the ROS level was not altered (Figure S5). In addition, we found that the antidepressant effects of NAC were reversed by the xCT inhibitor sulphasalazine (SAS) (Figure S6). These data support that NAC's effects are mediated through xCT and glutamatergic signaling modulation. Collectively, our data suggest that p38 MAPK phosphorylation and xCT downregulation contribute to CSDS-induced depressive-like behaviors in a KOR-dependent manner.
Additionally, we investigated the combined effects of norBNI with NAC through further experimentation. We tested lower doses of norBNI and NAC, both individually and in combination, on depressive-like behaviors in mice. Treatment with a low dose of norBNI or NAC alone did not significantly alleviate stress-induced depressive-like behaviors. However, the combined administration of norBNI and NAC effectively alleviated these behaviors (Figure S7), suggesting that the combination of norBNI and NAC was notably more potent than either agent used alone.
Following stress exposure, the mRNA level of xCT in the PL remained unchanged (Fig. 3A), suggesting that the posttranscriptional mechanism may contribute to CSDS-induced downregulation of xCT. MicroRNAs (miRNAs) are critical posttranscriptional regulators that inhibit translation or promote transcriptional repression, thus negatively regulating gene expression. Our previous miRNA-seq analysis identified several miRNAs with altered expression levels following CSDS [38]. Among these, miR-344d-1-5p and miR-3084-5p were predicted to interact with the 3'-UTR of xCT mRNA through bioinformatic analysis (Fig. 3B, C). We confirmed a significant increase in miR-344d-1-5p and miR-3084-5p expression following exposure to CSDS (Fig. 3D, E). Bioinformatic analysis suggested that the 3'-UTR of xCT mRNA contains putative binding sites for the seed-match sequences of miR-344d-1-5p and miR-3084-5p (Fig. 3F, G). To investigate these interactions, a dual-luciferase reporter assay was applied. HEK293 cells transfected with miR-3084-5p mimics showed a significant reduction in luciferase activity for the reporter containing the 3'-UTR of xCT, whereas mutant constructs had no significant effect (Fig. 3H, J). Transfection with miR-344d-1-5p mimics did not alter luciferase activity (Fig. 3I, J). Western blot analysis further demonstrated that xCT protein levels were reduced in cells infected with miR-3084-5p mimics, but not with miR-344d-1-5p mimics (Fig. 3K, L). These data support miR-3084-5p as a direct regulator of xCT. We further investigated whether modulation of KOR or p38 MAPK pathway affected miR-3084-5p expression. Administration of KOR antagonist norBNI decreased miR-3084-5p levels (Fig. 3M), and inhibiting the p38 MAPK pathway with SB203580 also reduced miR-3084-5p expression (Fig. 3N). These findings suggest that CSDS enhances miR-3084-5p expression and downregulates xCT in the PL through the activation of the KOR/p38 MAPK signaling pathway.
To determine the involvement of altered miRNA expression in CSDS-induced depressive-like behaviors, we designed oligonucleotides targeting miR-3084-5p and cloned them into an AAV vector (AAV-miR-3084-5p sponge) to suppress miR-3084-5p expression in the PL. As shown in Fig. 4A-D, intra-PL injection of the AAV-miR-3084-5p sponge resulted in a significant reduction of miR-3084-5p expression and increased xCT expression in the PL. Mice with reduced miR-3084-5p expression in the PL did not exhibit a significant depression phenotype after CSDS, as evidenced by increased social interaction in the SIT, decreased immobility in the TST and FST, and increased sucrose consumption in the SPT, without affecting locomotor activity (Fig. 4E-J). These findings indicate that downregulation of miR-3084-5p in the PL enhances xCT expression and alleviates depressive-like behaviors in mice. Together, these data support that increased miR-3084-5p expression plays a key role in xCT downregulation and CSDS-induced depressive-like behaviors.
To further validate the role of xCT, we overexpressed xCT in the PL via stereotactic injection of an AAV-expressing exogenous xCT (Fig. 4K). Enhanced xCT expression was confirmed by western blot analysis (Fig. 4L). Mice with xCT overexpression (AAV-EGFP-xCT) in the PL showed a significantly increased social interaction ratio in the SIT, reduced immobility time in the TST and FST, and increased sucrose consumption in the SPT after CSDS, without affecting locomotor activity (Fig. 4M-R). These findings underscore the critical role of the KOR-p38 MAPK-xCT pathway in CSDS-induced depressive-like behaviors.
Nasca with colleagues reported that stress can downregulate xCT expression [14], resulting in a failure to provide glutamate to prime nearby mGlu2 receptors. The resulting reduction in mGluR2-mediated inhibition leads to increased neuronal glutamate release, ultimately activating postsynaptic NMDA receptors. In alignment with these findings, we found that O-Acetyl-L-carnitine hydrochloride (LAC), known for its ability to epigenetically induce mGlu2 receptors, significantly alleviated CSDS-induced depressive-like behaviors (Figure S8A-F). Moreover, we observed that mGlu2/3 agonist LY379268 significantly alleviated CSDS-induced depressive-like behaviors (Figure S8G, H). Intra-PL injection of NMDA antagonist D-(-)-2-Amino-5-phosphonopentanoic acid (D-AP5) completely reversed stress-induced social avoidance (Figure S9). All together, these results support the critical role of xCT-mGlu2 interplay in CSDS-induced depressive-like behaviors.
Previous studies have predominantly demonstrated that KORs are expressed in various types of neurons and play a significant role in regulating depressive behaviors [32, 39, 40]. However, research has also indicated that KORs are endogenously expressed in striatal astrocytes, where they are capable of activating the p38 MAPK pathway [41, 42]. Given that xCT is predominantly expressed in astrocytes, we aimed to explore the involvement of KORs in prelimbic (PL) astrocytes in the CSDS-induced depressive-like behaviors. To investigate this, we first assessed the expression of KORs in PL astrocytes. Using in situ hybridization with immunofluorescence, we found that KORs were co-expressed with the astrocyte marker S-100β and the neuronal marker NeuN in the PL (Fig. 5A). The detailed analysis revealed that KOR was expressed in 65% of neurons and 15% of astrocytes (Fig. 5B). To investigate the function of astroglial KORs, we performed stereotaxic injection of AAV2/5-GfaABC1D-mCherry-T2A-Cre (GfaABC1D-mCherry-Cre) or AAV2/5-GfaABC1D-mCherry-WPRE (GfaABC1D-mCherry) into the PL of KOR mice, selectively knocking down KOR expression in astrocytes (Fig. 5C-E). The effectiveness of KOR knockdown was confirmed by immunofluorescence staining and in situ hybridization (Fig. 5F, G). We found that knockdown of KORs in astrocytes significantly alleviated CSDS-induced depressive-like behaviors, as indicated by increased social interaction in the SIT, enhanced sucrose consumption in the SPT, and reduced immobility time in both the TST and FST, without affecting locomotor activity (Fig. 5H-M). Astrocytic KOR knockdown also reversed CSDS-induced reduction in xCT expression (Fig. 5N). More importantly, we examined synaptic morphology following astrocyte-specific knockdown of KOR in the PL, and found that CSDS significantly reduced dendritic spine density, whereas astrocyte-selective KOR knockdown reversed this deficit (Fig. 5O, P). These data suggest that KORs in astrocytes play a crucial role in CSDS-induced depressive-like behaviors, providing new insights into the cellular mechanisms underlying stress-related depression.
To determine the role of miR-3084-5p and xCT in astrocytes, we microinjected AAV2/5-GfaABC1D-EGFP-miR-3084-5p sponge (GfaABC1D-miR-3084-5p sponge) or AAV2/5-GfaABC1D-EGFP-xCT (GfaABC1D-xCT) into the PL to selectively knock down miRNA-3084-5p or overexpress xCT in astrocytes (Fig. 6A-C, M-O). Viral reduction of miRNA-3084-5p or overexpression of xCT was confirmed by PCR or western blot (Fig. 6D, P). As shown in Fig. 6E-K, knockdown of miR-3084-5p in astrocytes increased xCT expression and alleviated CSDS-induced depressive-like behaviors, evidenced by a significant increase in social interaction in the SIT, enhanced sucrose consumption in the SPT, and decreased immobility time in the TST and FST, with no effect on the locomotor activity. Knockdown of miR-3084-5p in astrocytes also reversed CSDS-induced xCT reduction (Fig. 6L). Similarly, overexpression of xCT in astrocytes alleviated CSDS-induced depressive-like behaviors, without altering locomotor activity (Fig. 6Q-V). Moreover, astrocytic xCT overexpression also alleviated CSDS-reduced dendritic spine density (Fig. 6W, X). These data indicate that miRNA-3084-5p mediated xCT downregulation occurs in astrocytes, which plays a key role in CSDS-induced depressive-like behaviors.