Multiple brain regions manifest a discernible pattern of mGluR3 and mGluR5 expression in context of depressive-like behaviors induced by chronic stress
The selection of mice (Supplementary Fig. S1A, S1B), devoid of innate extreme motor and social behaviors, was conducted utilizing specific criteria21, including a social interaction (SI) test score ranging from 0.2 to 2, time spent in the center (TIC) during the open field test (OFT) exceeding 20 seconds, time spent in the open arms (TIO) of the elevated plus maze (EPM) exceeding 20 seconds, immobility time (IT) during the tail suspension test (TST) exceeding 20 seconds, and IT during the forced swim test (FST) exceeding 20 seconds. Subsequently, CSDS was administered to the selected cohort (27 control mice and 30 experimental mice). The outcomes revealed a significant reduction in SI scores and a notable increase in IT during both TST and FST in CSDS-exposed mice compared to control mice (Supplementary Fig. S1C, S1D). These alterations were evident not only when comparing the post-CSDS stage to the pre-CSDS stage within the CSDS group (Supplementary Fig. S1E, S1F) but also when contrasting the CSDS mice with the control mice (Fig. 1A and 1B). However, the duration of time spent in the central zone during the OFT and the time spent in the open arms of the EPM remained consistent between the two groups (Supplementary Fig. S1C, S1E, and Fig. 1A). These findings indicate that CSDS induces depressive behaviors rather than anxious behaviors in mice, rendering them suitable for further investigation.
Consequently, an analysis of mGluR3 and mGluR5 (dimer and monomer forms) expression was conducted in multiple brain regions associated with emotions, encompassing the medial prefrontal cortex (mPFC), hippocampus, amygdala, nucleus accumbens (NAc), bed nucleus of the stria terminalis (BNST), and ventral tegmental area (VTA) (Fig. 1C-H). Interestingly, compared to the control mice, the expression of mGluR3 was increased in the mPFC, amygdala, BNST, and VTA, while there was an elevation in mGluR5 protein levels in the hippocampus and BNST of the CSDS mice (Fig. 1C-H). These findings suggest that chronic stress influences the specific expression patterns of mGluR3 and mGluR5 in various emotion-related brain regions. However, the precise roles and expression patterns of mGluR3 and mGluR5 in different types of cells and their functions are still not fully understood.
In response to CSDS, astrocytes in mPFC and VTA exhibited mGluR3 upregulation, astrogliosis, and neuronal inhibition
Building upon prior findings and considering the prevalent expression of mGluR3 in astrocytes and the widespread expression of mGluR5 in neurons and astrocytes, our investigation focused on scrutinizing mGluR3 expression in astrocytes following CSDS. We utilized MACS to isolate astrocytes and validated their purity, achieving 90% (Fig. 2A). Notably, a substantial increase in mGluR3 expression (both dimer and monomer forms) was observed specifically in the isolated astrocytes from the medial prefrontal cortex (mPFC) of CSDS mice when compared to control mice (Fig. 2B). We also noted a significant expansion in the spreading area and an increased number of intersections between astrocytes at a distance of 50 µm from the cell soma (p = 0.052, Fig. 2C, 2D). Furthermore, an upregulation of multiple mRNA markers associated with reactive astrocytes, encompassing both pan-reactive (Lcn2, Timp1, CXcl10, Cd44, and Cp) and A1-specific (Serping1, Fbln5, Ugt1a, Fkbp5, and Srgn), was observed (Fig. 2E). Additionally, microdialysate collected from the mPFC of CSDS mice exhibited an increase in D-serine (p = 0.072), glutamate (p < 0.05), ATP (p = 0.17), and GABA (p = 0.16) levels compared to control mice (Fig. 2F, 2G). Remarkably, no detectable levels of mGluR3 were observed in the isolated mPFC microglia (Supplementary Fig. 2A, 2B), and there were no significant alterations in microglia activation-related genes (Supplementary Fig. 2C). In mPFC neurons (Supplementary Fig. 3A), CSDS mice exhibited similar levels of mGluR3 expression (Supplementary Fig. 3B) but a significant increase in the frequency of spontaneous inhibitory postsynaptic currents (sIPSC) and spontaneous excitatory postsynaptic currents (sEPSC) compared to control mice (Supplementary Fig. 3C-G). However, no significant changes were observed in the amplitude of sIPSC and sEPSC, as well as the frequency and amplitude of miniature IPSC (mIPSC) and miniature EPSC (mEPSC).
Furthermore, a similar upregulation astrocytes mGluR3 (dimer and monomer; Fig. 3A, 3B), astrocyte reactivation (Fig. 3C-G), and neuronal electrophysiological activities (Fig. 3H-L) but without affecting the mGluR3 level in neurons (Fig. 3H) and microglial activation (Supplementary Fig. 4A-C), were also observed in VTA while not in amygdala (Supplementary Fig. 5A-M) and BNST (Supplementary Fig. 6A-M).
mGluR3 deficiency ameliorated depressive-like behaviors in CSDS mice
To investigate the impact of mGluR3 on social and depressive-like behaviors (Supplementary Fig. 7A), we crossbred Grm3-KO mice with Thy1-EGFP mice, then screened and compared their behavior to wildtype (WT) mice in various tests (Supplementary Fig. 7B-F). Notably, CSDS induced depressive-like rather than anxiety-like behaviors in both WT (Supplementary Fig. 7C, 7D) and Grm3-KO mice (Supplementary Fig. 7E, 7F). Upon comparing Grm3-KO with WT mice post CSDS, no significant differences were observed in the time spent in center zones and open arms (Fig. 4A) while an increased SI ratio and reduced immobility time in the tail suspension and forced swim tests were evident (Fig. 4B).
To explore the changes in astrocytes in Grm3-KO mice after CSDS, isolated mPFC astrocytes (Fig. 4C) were examined. It was found that pan-reactive astrocytic reactivation (indicated by a decrease in Lcn2, S1pr3, Cxcl10, Cd44, and Vim) and A1-specific astrocytic reactivation (indicated by a reduction in H2.T23, Ggta1, Fbln5, Fkbp5, and Amigo2) were inhibited, but A2-specific astrocytic reactivation (indicated by an upregulation in sphk1 and B3gnt5) occurred compared to WT mice (Fig. 4D). Moreover, the spreading area of GFAP-positive astrocytes was significantly reduced, but not the interactions (Fig. 4E, 4F), as well as the secretion of gliotransmitters (P<0.05 for D-serine; P = 0.097 for glutamate; Fig. 4G). Evaluating the impact of gliotransmitters on neurons, it was observed that the frequencies of sIPSCs and sEPSCs were restored (Fig. 4H, 4I); however, the frequencies of mIPSCs and mEPSCs, as well as the amplitudes of all postsynaptic currents, remained similar between WT mice and Grm3-KO mice in the mPFC region (Fig. 4H-K).
Given the upregulation of mGluR3 expression on astrocytes and the decreased neuronal activity in the VTA region following CSDS, we also examined astrocytic reactivation and neuronal activity in VTA. In isolated VTA astrocytes (Supplementary Fig. 8A), the analysis revealed a decrease in pan-reactive and A1-specific genes from Grm3-KO mice compared to WT mice (Supplementary Fig. 8B). Additionally, results also showed that the spreading area of GFAP-positive astrocytes was increased (Supplementary Fig. 8C) while the concentrations of D-serine, glutamate, and GABA were decreased in the VTA (Supplementary Fig. 8D, 8E). Moreover, the frequency of sIPSCs and sEPSCs in the VTA region was increased, not in the sIPSCs and sEPSCs (Fig S8F-I).
In conclusion, the results of this study demonstrate that Grm3 knockout alleviates CSDS-induced social dysfunction and depressive-like behaviors, possibly through improving CSDS-induced changes in astrocytes reactivation (including morphology and neurotoxic transition), reducing excitatory gliotransmitter levels, and restoring CSDS-induced neuronal activity impairment. These findings highlight the role of mGluR3 in modulating astrocyte-neuron interactions and provide insights into potential therapeutic targets for social and depressive disorders.
Therapeutic Potential of Inhibiting mGluR3 Activity: Alleviating Depressive Behaviors and Restoring Neural Electrophysiological Activities in Mice
To explore the potential of therapeutic interventions of by targeting mGluR3 for depression, FLIPR assays were conducted to screen the impact of mGluR3 inhibition using glutamic acid, MTPG, DHPG, and LAP4 on various cell types. As anticipated, no calcium influx was observed in HEK-293T cells; however, significant calcium flux was induced in HEK cells expressing mGluR3 when exposed to glutamic acid and MTPG, with no effect observed for DHPG or LAP4. The agonistic effects of glutamic acid and MTPG were further validated in primary astrocytes, where glutamic acid exhibited a lower EC50 than MTPG, indicative of specific agonism on mGluR3 (Supplementary Fig. 9A). To confirm the specific antagonistic effect of NAM (VU786) on mGluR3, Grm3-HEK cells, Grm5-HEK cells, primary astrocytes, and IPSCs-astrocytes were pre-treated with NAM or PBS before exposing them to glutamic acid (Supplementary Fig. 9B, 9C) or MTPG (Supplementary Fig. 9D, 9E) at various concentrations. The results demonstrated that NAM selectively inhibited the agonistic effects of glutamic acid and MTPG induced calcium influx on mGluR3 in Grm3-HEK cells, primary astrocytes and IPSC induced astrocytes.
In term of behavioral screening on 2-month Thy1-EGFP mice, those with normal social and mood functions were selected and treated with an mGluR3-NAM enriched diet (29 Thy1-EGFP mice) or normal diet (28 Thy1-EGFP mice) from day 60 to day 104 (Supplementary Fig. 10A, 10B). Confirming no difference in anxiety-like behaviors between the Normal diet group and NAM treated group after CSDS, significant social dysfunction and depressive-like behaviors were demonstrated in both groups after CSDS (Supplementary Fig. 10C-F). NAM-treated mice exhibited similar anxiety-like and social behavior (Fig. 5A, 5B) while showing a marked reduction in immobility time during the FST compared to normal diet mice after CSDS, with a slight decrease in immobility time during the TST (P = 0.088, Fig. 5B). In isolated mPFC astrocytes (Fig. 5C), mRNA expression specific for pan-reactive astrocytes (Lcn2, Cxcl10, Cd44, Cp, Serpina3n, Aspg) and A1 specific astrocytes (H2.T23, Serping1, Ggta1, Fbln5, Fkbp5, Amigo2) was decreased, while the mRNA level of Emp1 specific for A2 phenotypic astrocytes was increased (Fig. 5D). Morphological visualization data showed that NAM treatment inhibited astrocytic hypertrophy by reducing the spreading area of GFAP-positive astrocytes; as well as D-serine and glutamate levels (P<0.05 for D-serine; P = 0.097 for Glutamate; Fig. 5G). Additionally, electrophysiological assay revealed enhanced frequencies of sIPSC and sEPSC in NAM-treated mice compared to normal diet mice (Supplementary Fig. 11A-D).
Simultaneously, in VTA astrocytes (Supplementary Fig. 12A), mRNA expression specific for pan-reactive (Lcn2, Steap4, S1pr3, Timp1, Cp, Vim) and A1 astrocytes (H2.T23, Serping1, H2.D1, Fbln5, Psmb8, Amigo2) was decreased, while Emp1 mRNA expression was increased (Supplementary Fig. 12B); in parallel with the reduction of spreading area and number of astrocyte intersections (Supplementary Fig. 12C, 12D), and concentrations of D-serine and glutamate (Supplementary Fig. 12E). Conversely, NAM treatment increased the frequencies of sIPSC and sEPSC in the VTA region compared to normal diet mice (Supplementary Fig. 12F-I).
In summary, these findings demonstrate that NAM treatment alleviates astrocytes reactivation by reducing morphological and neurotoxic phenotypic alterations, gliotransmitter release, and enhancing neural activity in mPFC and VTA regions.
Selective Astrocytes mGluR3 inhibition in mPFC and VTA alleviated depressive behaviors, neurotoxic reactivation, and neural activity dysfunction
In order to further investigate the significance of astrocytes mGluR3 in mPFC and VTA, we specifically injected AAV5-GfaABC1D-mCherry (25 control-AAV) or AAV5-GfaABC1D-Cre-mCherry (27 knockdown-AAV) into the mPFC and VTA of behavioral screened 3-month-old Grm3-Flox (Thy1-EGFP) mice (Supplementary Fig. 13A, 13B) to specifically knockdown astrocytes mGluR3. As expected, CSDS induced distinct social dysfunction and depressive-like behaviors in both control-AAV (Supplementary Fig. 13C, 13D) and knockdown-AAV injected mice (Supplementary Fig. 13E, 13F). When comparing control-AAV mice with knockdown-AAV mice, specific astrocyte mGluR3 knockdown (Fig. 6A, Supplementary Fig. 14A-D) lessen social dysfunction and depressive-like behaviors nearly approaching significance (p = 0.081 in the SI test; p = 0.19 in the TST; p = 0.12 in the FST; Supplementary Fig. 13E, 13F). Furthermore, spreading area and number of astrocyte intersections, mRNA expression of pan-reactive and A1 phenotypic reactivation markers, the release of gliotransmitters were all reduced in both mPFC (Fig. 6D-F) and VTA (Fig. 6G-I) of knockdown AAV injected mice. Electrophysiological investigation also revealed that the frequencies of sIPSC and sEPSC was reversed in mPFC (Supplementary Fig. 14E-I) but not in VTA (Supplementary Fig. 15A-E). These results indicate that selectively astrocytes mGluR3 knockdown in mPFC and VTA enhances resilience to CSDS-induced social dysfunction and depressive-like behaviors in mice.