The connexin hemichannel inhibitor D4 produces rapid antidepressant-like effects in mice

Depression is a common mood disorder characterized by a range of clinical symptoms, including prolonged low mood and diminished interest. Although many clinical and animal studies have provided significant insights into the pathophysiology of depression, current treatment strategies are not sufficient to manage this disorder. It has been suggested that connexin (Cx)-based hemichannels are candidates for depression intervention by modifying the state of neuroinflammation. In this study, we investigated the antidepressant-like effect of a recently discovered selective Cx hemichannel inhibitor, a small organic molecule called D4. We first showed that D4 reduced hemichannel activity following systemic inflammation after LPS injections. Next, we found that D4 treatment prevented LPS-induced inflammatory response and depressive-like behaviors. These behavioral effects were accompanied by reduced astrocytic activation and hemichannel activity in depressive-like mice induced by repeated low-dose LPS challenges. D4 treatment also reverses depressive-like symptoms in mice subjected to chronic restraint stress (CRS). To test whether D4 broadly affected neural activity, we measured c-Fos expression in depression-related brain regions and found a reduction in c-Fos+ cells in different brain regions. D4 significantly normalized CRS-induced hypoactivation in several brain regions, including the hippocampus, entorhinal cortex, and lateral septum. Together, these results indicate that blocking Cx hemichannels using D4 can normalize neuronal activity and reduce depressive-like symptoms in mice by reducing neuroinflammation. Our work provides evidence of the antidepressant-like effect of D4 and supports glial Cx hemichannels as potential therapeutic targets for depression. Supplementary Information The online version contains supplementary material available at 10.1186/s12974-023-02873-z.


Introduction
Major depression is one of the most prevalent mood disorders worldwide.Individuals with depression display a variety of symptoms marked by prolonged depressed mood and diminished interest in activities [1,2].Although decades of clinical and basic research have provided significant insights into the pathophysiology of depression, the underlying neural basis remains unclear [3,4].Current treatment strategies are not sufficient to alleviate the burden caused by depression.For example, classic antidepressant drugs targeting traditional neurotransmitter systems, like serotonin, have delayed onsets [5].Novel fast-acting antidepressants, such as ketamine, offer rapid relief from depressive symptoms but face restricted clinical use due to adverse effects, including safety in long-term treatment [6].The future development of antidepressant medications still requires a better understanding of the neural mechanisms involved in depression and the identification of effective molecular targets.
In the central nervous system, hemichannels formed by connexins (Cxs) or pannexins are essential for neuronglia communication and maintenance of brain homeostatic balance [7,8].In normal conditions, glial Cx-based hemichannels have relatively low activity.Under physiological conditions, neuronal pannexin hemichannels are permeable to small molecules like ATP and are preferentially permeable to Cl - [9,10].However, exacerbated neuronal hemichannel activity has been observed in diverse pathological conditions [11].Under these conditions, hemichannel activity increases and promotes the release of various gliotransmitters, including ATP and glutamate [7].Hemichannel hyperactivity can alter neuronal excitability and thus promote neuronal excitotoxicity, neuroinflammation, and cell damage in several brain diseases, such as epilepsy and ischemia [12][13][14][15][16][17].While recent studies have shown that some hemichannel blockers can provide neuroprotective effects [16,18], the therapeutic potential of hemichannel regulation has been scarcely investigated.
Downregulation of connexin 43 (Cx43) in the prefrontal cortex of patients with depression [19][20][21] has been partially validated in an animal model of depression [22].Interestingly, the blockade of Cx-based channels (hemichannels and gap junction channels) potentiates the effect of antidepressant treatment [23].Moreover, mounting evidence supports an inhibitory impact of antidepressants on Cx hemichannels.For instance, antidepressant suppresses Cx hemichannel activity induced by LPS in cultured cortical astrocytes [24].Another study indicated that glutamate release from activated Cx hemichannels could be inhibited with antidepressant drugs [25].Ketamine, a recently approved fast-acting antidepressant [26], can inhibit hemichannel activity in cortical astrocytes and neurons induced by LPS or proinflammatory cytokines [27].These findings suggest that hemichannels may serve as a promising target for antidepressant treatments.However, although recent clinical and preclinical findings suggest a role for Cx hemichannels in depression [28], in vivo results supporting the role of hemichannels in alleviating depressive-like symptoms are scarce.
In a recent study, we screened and identified a novel Cx hemichannel inhibitor, which is a small organic molecule called D4 that has been shown to be effective in reducing the atrophy in pathological muscular dysfunction [29] as well as alleviating neuroinflammation in temporal lobe epilepsy [16].Here, we hypothesize that blocking Cx hemichannels can alleviate neuroinflammatory and depressive-like symptoms.Using a combination of dye uptake experiments, immunofluorescent staining of astrocytes and microglia, immediate early gene c-Fos expression, and behavioral analysis, we investigated the antidepressant-like efficacy of D4.We found that blocking mainly Cx hemichannels with D4 produces antidepressant-like effects by suppressing neuroinflammation.

Animals
Adult C57BL/6 mice (2-4 months old) were used in this study.Male mice were used throughout this study except for data presented in Figs. 2 and 4, where both male and female mice were used.Mice were acclimated to the Laboratory Animal Research Unit (LARU) at the City University of Hong Kong.Mice were group housed in a 12 h dark/light cycle (light on/off at 08:00/20:00).Food and water were given ad libitum.All animal procedures were approved by the LARU and were conducted in accordance with guidelines from the Animal Research Ethics Sub-Committees of the City University of Hong Kong and the Department of Health of the Government of the Hong Kong Special Administrative Region.

CBF dye loading
Hemichannel opening in brain slices was assessed by CBF (5(6)-Carboxyfluorescein, 21877, Sigma-Aldrich) dye uptake as previously published [30].Briefly, acute coronal hippocampal slices (300 µm) were prepared in ice-cold oxygenated (95% O 2 and 5% CO 2 ) ACSF containing (in mM) 119 NaCl, 2.5 KCl, 2.5 CaCl 2 , 1.3 MgSO 4 , 1 NaH 2 PO 4 , 26.2 NaHCO 3 , and 22 glucose.For the dye loading experiments after acute LPS challenges, acute slices were transferred to ACSF (vehicle) or D4 and incubated for 20 min.CBF (100 µM) was added to ACSF saturated with 95% O 2 and 5% CO 2 at room temperature.After pre-treatment, slices were then transferred to ACSF containing CBF for 20 min.For the dye loading experiments after repeated LPS challenges or CRS, acute slices were incubated with ACSF saturated with 95% O 2 and 5% CO 2 at room temperature for 20 min.Then, incubated slices were transferred to ACSF containing CBF (100 µM) saturated with 95% O 2 and 5% CO 2 at room temperature for 20 min.Slices were washed three times with ACSF and fixed with 4% paraformaldehyde at 4 °C overnight.Slices were cryoprotected in 30% sucrose for 36 h before being fast frozen, followed by cryosectioning.Brain sections (15 µm thick) were prepared with a cryostat (HM525 NX, Thermo Fisher Scientific) for immunostaining and confocal microscopy.Fluorescence images were quantified using ImageJ (National Institutes of Health).Briefly, background subtraction was applied to the whole image to reduce non-specific fluorescence signals of raw images.Quantitative analysis was performed on thresholded images to analyze the area or integrated density (IntDen) of positive signals.For a given marker, the threshold value and size for particle analysis were held constant for each set of experiments.To analyze CBF uptake in different neural cells, thresholded images of the channel containing either glial fibrillary acidic protein (GFAP), allograft inflammatory factor 1 (Iba1), or Nissl were first used to extract regions of interest (ROIs) representing astrocyte, microglia, or neuronal somata, respectively.Then, CBF uptake in each cell type was measured by quantifying CBF-positive areas of the selected ROIs.

Mouse models of depression
Inflammation-based mouse model of depression: Systemic inflammation activated by LPS (L2880, Sigma-Aldrich) can induce depressive-like behaviors in mice.The LPS-induced depression model was conducted as previously described [31].LPS was dissolved in sterile 0.9% saline.Mice were intraperitoneally (i.p.) injected with LPS or saline at a 0.75 mg/kg dose between 10:00 to 12:00 daily for 1 week.Body weight was measured before each injection.Behavioral assays were performed 24 h after the last injection.
Stress-induced mouse model of depression: The CRS-induced depression model was used as previously described [32].Mice were restrained in the ventilated 50 mL Falcon tubes (with sixteen 3.5 mm air vents at the wall and one 2.0 mm air vent at the nasal end of the tube) for 6 h per day during the light cycle.Mice were able to move their head and body but could not escape.During the restraint, animals had no access to food/water or social interaction.As a control, a separate cohort of mice was subjected to food and water deprivation for 6 h.Body weight was measured once a week.Behavioral assays were performed 24 h after the last restraint.

Behavioral assays
All behavioral tests were performed during the light cycle.Mice were habituated to the behavior room for at least 1 h before testing.ToxTrac software was used to analyze the behavioral videotape offline [33].The behavior box was cleaned with 70% ethanol after each trial to eliminate any olfactory cues.
Open field test (OFT): Mice were randomly placed in one corner of the open-field arena (50 cm × 50 cm × 40 cm, Length × Width × Height) with light (400 lx) and were allowed to explore freely for 10 min.A top camera above the arena was used to record mouse movement.The total distance and time spent in the center of the arena (25 cm × 25 cm) were analyzed using Toxtrac software by tracking the mouse centroid.The total distance indicated the locomotor and exploratory activities of mice.Less time spent in the center suggests anxiety-like behavior [34].
Tail suspension test (TST): Mice were suspended by their tails and secured with tape for 6 min in a behavior box with dim light (100 lx).Each mouse was separated into its three-walled area without visual interaction.A 3D-printed white hollow tube was used to prevent climbing during the test.A side camera was placed in the behavior box to record the mouse's movement.The mouse speed was measured by tracking the mouse centroid using Toxtrac software.Total immobility time (speed ≤ 0.5 cm/s) or struggling time (speed > 5 cm/s) was counted during the last 4 min.Increased immobility or decreased struggling suggests depressive-like behavior [35].
Forced swim test (FST): Mice were introduced to a cylindrical container (22 cm inner diameter, 20 cm depth) filled with 10 cm-deep tap water (22-24 °C) and allowed to swim for 6 min.A top camera above the behavior box was used to record the mouse's movement.Mouse speed was analyzed by tracking the mouse centroid using Toxtrac software.Total immobility time (speed ≤ 1.25 cm/s) was counted during the last 4 min.An increase in immobility suggests depressive-like behavior [36].
Sucrose preference test (SPT): Mice were habituated for 48 h to 1% sucrose water, followed by 24 h of water deprivation.During the test day, mice were placed in the behavior box with a pre-weighed bottle filled with 1% sucrose or plain water for 1 h to determine their preference for 1% sucrose or water.Bottles were weighed after the test.Sucrose preference was expressed as (sucrose intake)/(sucrose intake + water intake) × 100.Reduced sucrose preference suggests anhedonia [37].

Immunofluorescence and analysis
After behavioral tests, animals were deeply anesthetized with a 10% ketamine/1% xylazine mixture and were then perfused with 4% paraformaldehyde (PFA) in 1X phosphate-buffered saline (PBS) at room temperature for 10 min.The whole brain was post-fixed overnight in 4% PFA at 4 °C followed by cryoprotection in 30% sucrose in PBS at 4 °C for 72 h.Brain sections (30 µm thick) were prepared with a cryostat (HM525 NX, Thermo Fisher Scientific).For immunohistochemistry, the sections were washed three times with 1X PBS for 10 min each, followed by 10% normal goat serum in 1X PBS.Preparations with primary antibodies (rabbit anti-GFAP, 1:500, Z0334, Dako; rabbit anti-Iba1, 1:500, 019-19741, Wako; rabbit anti-c-Fos, 1:500, ab190289, Abcam) were applied for overnight at 4 °C.Next, sections were incubated with the secondary antibody (Jackson ImmunoResearch) with 4′,6-diamidino-2-phenylindole (DAPI, 28718-90-3, Santa Cruz Biotechnology) and/or Nissl stain (NeuroTrace 640/660, N21483, Thermo Fisher Scientific) at room temperature for 2 h in darkness.After washing with 1X PBS, sections were mounted (Antifade mounting medium, H-1000, Vector Laboratories).The stained sections were stored in the dark box at 4 °C before imaging.Images were visualized by epifluorescence (Nikon Eclipse Ni-E upright fluorescence microscope, Nikon) or confocal microscopy (LSM880, ZEISS).Fluorescence images were quantified using ImageJ (National Institutes of Health) as described above.Briefly, the raw images were processed with background subtraction.To quantify the number of positive cells, the threshold value and size for particle analysis were manually adjusted for each marker.

Cytokine analysis
Mouse blood samples and hippocampi were harvested 4 h after the last LPS injection and drug treatment.To collect plasma, blood was collected into tubes containing ethylenediaminetetraacetic acid (EDTA) and chilled on ice before centrifugation.Plasma was separated by refrigerated centrifugation at 2000×g for 20 min (4 °C) within 1 h of collection.After centrifugation, the plasma was immediately extracted and then frozen at − 20 °C before use.For protein extraction, hippocampi were dissected on ice and rinsed three times with ice-cold 1X PBS.Then, hippocampi were lysed in the radioimmunoprecipitation assay (RIPA) buffer with a protease inhibitor cocktail (1:200), sodium pyrophosphate (1 mM), and 20 mM sodium fluoride (Sigma-Aldrich).The hippocampal lysates were centrifuged at 12,000 rpm for 15 min at 4 °C to collect the supernatant.Protein concentrations were determined with DC protein assay kit (5000111, Bio-Rad).Plasma and hippocampal interleukin-1β (IL-1β) protein levels were measured by an enzyme-linked immunosorbent assay (ELISA) (EK0394, Boster Biological Technology) according to the manufacturer's instructions by fluorescence measurement at 450 nm (Synergy H1 Microplate Reader, BioTek Instruments).The level of IL-1β was expressed as fold changes compared to the saline/vehicle control group average.

Statistics
Statistical analysis was performed using SPSS 25 (International Business Machines Corporation, IBM).For comparisons between data with normal distributions and equal variances, one-way analysis of variance (ANOVA) was used to check differences between multiple groups.Student's t-test was performed to test differences between two groups.A nonparametric test was used when data did not pass tests for normality and equal variance.Unless otherwise specified, data are shown as mean ± s.e.m.Thresholds for significance were indicated as *p < 0.05, **p < 0.01, and ***p < 0.001.The details of statistical tests and results are shown in the figure legends.

D4 suppresses LPS-induced astrocytic activation and hemichannel activity
Reactive glial cells are hallmarks of neuroinflammation and have been implicated in the pathophysiology of depression [52][53][54][55].Having demonstrated the antidepressant-like properties of D4, we first investigated the effect of D4 on LPS-mediated inflammation.We injected adult mice intraperitoneally with a low dose of LPS (0.75 mg/kg, i.p.) or saline (control) daily for 1 week.Mice were orally fed with either D4 (5 mg/ kg, p.o.) or vehicle after each LPS injection.Four hours after the last LPS injection, plasma and hippocampi were extracted for ELISA assay.We found that repeated LPS injections significantly increased pro-inflammatory cytokine IL-1β levels in the plasma and hippocampi.
Together, these data suggest that D4 can preferentially block astrocytic Cx hemichannels, and that LPS-induced increases in membrane permeability of microglia might be mediated by other pathways not inhibited by D4, such as pannexin1 hemichannels [7].
Next, we compared depressive-like behavior using the depressive-like z scores.Oral administration of D4 in a different treatment paradigm did not affect depressivelike z scores in the unstressed control mice (Vehicle, 0.00 ± 0.12; D4, − 0.22 ± 0.15) (Fig. 4G).Indeed, mice exposed to CRS exhibited significantly higher depressivelike z scores than the control/vehicle group.The CRSinduced depressive-like phenotype in mice was again ameliorated by D4 (Control/vehicle, 0.00 ± 0.14; CRS/ vehicle, 1.02 ± 0.15; CRS/D4, 0.22 ± 0.13) (Fig. 4N).Given that two doses of D4 were administered before the beneficial effects in TST were observed, the earliest positive effects exerted by D4 were approximately two days.These results could consolidate the antidepressant-like effects of D4 and demonstrate its fast-acting potential using a different mouse model of depression.

D4 restores neuronal activity in CRS-induced depressive-like mice
Alterations in neural activity and brain network connectivity are consistently associated with depression [59][60][61][62].Glial cells are important for supporting the functions of neurons [63,64].In addition, astroglial Cx hemichannels have been shown to regulate behavior by modulating neuronal activity [65,66].Therefore, we asked whether the antidepressant-like effects of hemichannel blockade with D4 could be accompanied by changes in neural activity.
Subsequently, we assessed neuronal activation using c-Fos immunostaining following acute exposure to stress (TST).In mice subjected to CRS, there were significantly fewer c-Fos + cells in several depression-related brain regions, including the dorsal and ventral hippocampus (dDG, vDG, vCA1, vSub), ventral entorhinal cortex, lateral septum, and nucleus accumbens (NAc).CRS-exposed mice treated with D4 (5 mg/kg, p.o.) exhibited increases in the number of c-Fos + cells in the hippocampus (dDG, vDG, vCA1), entorhinal cortex, and lateral septum.No significant change in the number of c-Fos + cells was found in the amygdala (Fig. 6E, F).These data suggest that D4 may improve behavioral deficits in depressive-like mice by restoring neural activity in various brain regions.

Discussion
Hemichannels are important for regulating cellular communication and maintaining homeostasis [8].Uncontrolled hemichannel activation has been proposed to disrupt neuron-glia communication, promote inflammation, and lead to cell damage in various brain diseases [7].Recent evidence suggests that elevated hemichannel activity is associated with pathological changes in depression [67].Therefore, inhibiting hemichannel activity may provide therapeutic benefits, as previously suggested [67][68][69].
Accordingly, in the present study, we demonstrated the rapid, antidepressant-like effects of D4, which reduced depressive-like symptoms in mice exposed to repeated systemic LPS challenges or CRS.We also showed that blocking hemichannels with D4 inhibits stress-mediated reactive astrogliosis and induces a brain-wide restoration of neural activity in the depressive-like mice subjected to CRS, which may contribute to the behavioral effects of D4.Together, our findings indicate that Cx hemichannels can constitute effective therapeutic targets for depression, and that D4 can be a new molecule mediating this therapeutic effect.
The pathogenesis of depression is highly heterogeneous and is associated with intricate interplays across multiple neurotransmitter systems, neuroimmune systems, neural circuity, and brain networks [70][71][72].Considering that neuroinflammatory responses are likely to affect the ontogeny of the central nervous system [73], outcomes under these conditions could be irreversible, and could generate a radical difference compared to the effects of neuroinflammation that only occurs in adulthood.These differences limit current treatment outcomes and dramatically impede the rational design of successful therapies.
Cx hemichannels have been suggested as potential targets for depression [24,67,74].Several agents, including small molecules, antibodies, and peptides, are available for inhibiting Cx hemichannels, with most studies carried out in vitro [18].However, the behavioral benefits of hemichannel inhibition in depression in vivo remain elusive.In the present study, using two well-established mouse models of depression, we provide substantive evidence for the first time that D4 treatment can improve stress-induced depressive-like symptoms in adult mice.In addition, adult mice that received either acute or chronic applications of D4 did not show evident side effects.The latter could suggest that D4 is a relatively safe Cx hemichannel inhibitor with therapeutic effects.The antidepressant-like potential of D4 still needs to be substantiated for future translational research and clinical studies.Future work is also required to determine its bioavailability and druggability.
It has been shown that hemichannel activity can increase due to inflammation and chronic stress in adult animals [39,67].One of the prevailing hypotheses is that abnormal hemichannel activity may amplify detrimental effects on several cellular processes and promote the pathogenesis of a range of neurological and neurodegenerative diseases [10,16,43,75,76].For example, in animal models of brain injury, such as ischemic stroke, hemichannels are activated in response to injury and pro-inflammatory cytokines.Dysregulated hemichannel activity can facilitate excitotoxicity due to the excessive release of gliotransmitters and may lead to neuronal death [77].During epileptogenesis, seizures increase glial hemichannel activity [16,78].In turn, such heightened hemichannel activity promotes seizure generation and propagation through neuroinflammation and neuronal hyperexcitability [15,79].In animal studies, currently available hemichannel inhibitors have been used to control diseases like ischemia and epilepsy [12,15].Consistent with a previous study [41], we found that a single high-dose LPS injection increases glial hemichannel activity in the hippocampus.In addition, we found that repeated low doses of LPS challenges can increase Cx hemichannel activity in the hippocampal DG.Blocking hemichannels with D4 also prevented LPS-mediated reactive astrogliosis, probably due to reduced hemichannel activity in astrocytes, as evidenced by decreased CBF dye uptake.Thus, decreases in Cx hemichannel activity may contribute to antidepressant-like behavioral effects exerted by D4 in the LPS-induced depression model.These findings are consistent with the hypothesis that inhibition of Cx hemichannels by small molecules or other agents can improve depressive-like behaviors.
Our finding that blocking Cx hemichannels with D4 restored neuronal activity in CRS-exposed depressivelike mice provides another mechanistic insight into how hemichannel inhibition can contribute to behavioral improvement.Reduced neuronal activity in several regions of the limbic system has been identified in the etiology of depression [80].Notably, post-mortem and neuroimaging studies have confirmed that the hippocampus is hypoactive in patients with depression [81,82].Antidepressant treatments can reverse depressionassociated structural and functional changes in the hippocampus [83][84][85].
Similar to previous findings, we found that mice acutely exposed to CRS display significant reductions in neuronal activity in the hippocampus [86] and several other limbic regions, including the entorhinal cortex and the lateral septum.In contrast, D4 treatment reduces depressive-like behavior in the TST and enhances neural activation, as evidenced by the recovery of c-Fos + cells in these regions.Although our dye uptake and glial density assays were restricted to the hippocampal DG regions, findings from our c-Fos screening results raise the possibility that D4 could impact the hemichannel and neural activity in multiple sites across the brain after systemic application via oral administration.Evidence for this possibility is supported by our recent study showing that oral treatment of D4 with optimized doses is sufficient to reduce seizure-induced glial cell density and hemichannel activity in different subfields of the hippocampus and anterior piriform cortex [16].Data from clinical and preclinical studies suggest that the route of antidepressant administration may affect therapeutic efficacy and experience, especially for the novel rapid-acting antidepressant such as ketamine [6,85].Thus, specific mechanisms by which CRS reduces neuronal activation and D4 regulates neural activity remain to be studied with more sophisticated and precise targeted drug delivery strategies.
Cx43 is predominantly expressed in astrocytes in the adult mouse brain.Resting microglia rarely express Cx43, but its expression level can be increased in activated microglia [75,87,88].Previously, it has been shown that D4 inhibits Cx43 hemichannels but not gap junction channels [29].In this study, we showed that D4 reduces LPS-induced CBF uptake in astrocytes without affecting dye uptake in microglia.D4-induced reduction in astroglia-mediated hemichannel activity could indirectly reduce pannexin1 hemichannel activity in neurons, as observed under neuroinflammation triggered by a neurotoxic beta-amyloid peptide [89].Based on existing studies suggesting that hemichannels can regulate gliotransmitter release, synaptic transmission, and neuronal firing [90][91][92][93][94], we speculate that the behavioral effect of D4 is likely mediated by the initial inhibition of astrocytic Cx43 hemichannel activity, which subsequently affects neuron-glia interaction and neuronal function.A limitation of our current study was the lack of detailed comparisons between sexes in their sensitivity to chronic stressors and compound D4.Sex could be a factor that affects the prognosis of depression and the treatment response of antidepressant interventions [2,5].It will be interesting to test whether D4 has a sex-specific effect on changes in hemichannel and behavioral activity induced by depressogenic factors.Given that Cx hemichannels can actively regulate neural activity and behavioral states via a cornucopia of cellular processes [64,65], future studies are needed to elucidate the mechanisms of antidepressantlike behavioral effects of D4.

Conclusion
In conclusion, we have shown that a recently discovered hemichannel inhibitor, namely the small organic molecule D4, can exert antidepressant-like effects in mice subjected to repeated systemic LPS challenges or CRS.These behavioral benefits of D4 are accompanied by the blockade of hemichannel activity reduced inflammatory response, and a brain-wide restoration of neural activity in depressive-like mice.Our findings support the hypothesis that the hemichannel inhibitor is a promising pharmacophore for future antidepressant development focusing on glial cells.A better understanding of D4's cellular mechanisms may help to identify its therapeutic potential and promote clinical uses of hemichannel inhibitors in mood disorders.

(
See figure on next page.)Fig. 1 D4 inhibits LPS-induced hemichannel activity.A Experimental design to examine CBF uptake after acute systemic bacterial lipopolysaccharide (LPS) injection.B Representative images showing the CBF uptake in the acute hippocampal dentate gyrus (DG).CBF, green; DAPI, blue.Scale bar: 50 µm.C Quantification of CBF + area in the stained DG sections.D Normalized CBF + integrated density (IntDen) in the DG sections of saline/vehicle control (Ctrl), LPS/vehicle, LPS/D4 (0.1 µM), LPS/D4 (1 µM), and LPS/D4 (10 µM).The concentration of ethanol in D4 solution is no more than 0.02% v/v.n = 15 DG sections from 5 mice for all groups.E Protocol to measure CBF uptake after repeated systemic low-dose LPS injections.F Representative images showing the CBF uptake in the DG of mice received vehicle (Veh) or low-dose D4 (0.5 mg/kg, p.o.).CBF, green; DAPI, blue.Scale bar: 50 µm.G Violin plot of normalized CBF + area in the DG of LPS/vehicle and LPS/D4 (0.5 mg/kg, p.o.) treated mice.H Violin plot of normalized CBF + integrated density in the DG.Saline/vehicle, white, n = 9 DG sections from 3 mice; LPS/vehicle, red, n = 18 DG sections from 7 mice; LPS/D4 (0.5 mg/kg, p.o.), blue, n = 19 DG sections from 7 mice.I Representative images showing the CBF uptake in the DG of mice received vehicle (Veh) or higher dose D4 (5.0 mg/kg, p.o.).J Violin plot of normalized CBF + area in the DG.K Violin plot of normalized CBF + integrated density in the DG of LPS/vehicle and LPS/D4 (5.0 mg/kg, p.o.) treated mice.Saline/vehicle, white, n = 8 DG sections from 3 mice; LPS/ vehicle, red, n = 11 DG sections from 5 mice; LPS/D4 (5.0 mg/kg, p.o.), blue, n = 10 DG sections from 4 mice.The open circle shows the value of each DG.Kruskal-Wallis test with Bonferroni correction (

Fig. 2
Fig. 2 D4 prevents depressive-like behaviors induced by repeated low-dose LPS challenges.A Schematic of drug treatment and behavioral testing.B Center time and C total distance in the OFT.D Time immobile in the TST.E Time immobile in the FST.F Sucrose preference in the SPT.G Depressive-like z score of the unstressed mice treated with vehicle or D4.Ctrl/vehicle, n = 10 mice, 5 male and 5 female; Ctrl/D4, n = 10 mice, 5 male and 5 female.H Timeline of LPS injection, drug treatment, and behavioral testing.I Center time and J total distance in the OFT for saline/vehicle control, LPS/vehicle treated, and LPS/D4 treated mice.K Time immobile in the TST.L Time immobile in the FST.M Sucrose preference in the SPT.N Depressive-like z score of saline/vehicle control, LPS/vehicle, and LPS/D4 mice.Saline/vehicle, n = 12 mice, 7 male and 5 female; LPS/vehicle, n = 21 mice, 15 male and 6 female; LPS/D4, n = 21 mice, 12 male and 9 female.The filled dot indicates the value of each mouse.Data are mean ± s.e.m.Unpaired Student's t-test (B-G).One-way ANOVA followed by Fisher's Least Significant Difference (LSD) post hoc test (J, K, M, N).Kruskal-Wallis test with Bonferroni correction (I, L). *p < 0.05, ***p < 0.001

Fig. 4
Fig. 4 D4 improves depressive-like behaviors induced by chronic restraint stress (CRS).A Schematic of drug treatment and behavioral testing.B Center time and (C) total distance in the OFT.D Time immobile in the TST.E Time immobile in the FST.F Sucrose preference in the SPT.G Depressive-like z score of the unstressed mice treated with vehicle or D4.n = 13 mice (5 male and 8 female) for each group.H Schematic of the design of CRS, drug treatment, and behavioral testing.I Center time and J total distance in the OFT.K Time immobile in the TST.L Time immobile in the FST.M Sucrose preference in the SPT.N Normalized depressive-like z score of control/vehicle, CRS/vehicle, and CRS/D4 treated mice.Control/vehicle, n = 10 mice, 5 male and 5 female; CRS/vehicle, n = 18 mice, 9 male and 9 female; CRS/D4, n = 19 mice, 10 male and 9 female.The filled dot indicates the value of each mouse.Data are mean ± s.e.m.Unpaired Student's t-test (B, C, E-G).Mann-Whitney U test (D).One-way ANOVA followed by LSD post hoc test (L-N).Kruskal-Wallis test with Bonferroni correction (I-K).*p < 0.05, **p < 0.01, ***p < 0.001

Fig. 6
Fig. 6 D4 rescues hypofunction of depression-related brain regions in the CRS-induced depressive-like mice.A Schematic of the design of CRS, drug treatment, behavioral testing, and c-Fos immunostaining after tail suspension test (TST).B Depressive-like z score of mice used for the TST and c-Fos experiments.C Time immobile and struggling (D) in the TST.The filled dot shows the value of each mouse.Ctrl/vehicle, white; CRS/ vehicle, red; CRS/D4, blue.n = 5 mice for all groups.Data are mean ± s.e.m.One-way ANOVA followed by LSD post hoc test (B).Kruskal-Wallis test with Bonferroni correction (C, D).E Immunostaining for c-Fos expression (red) in the depression-associated brain regions of ctrl/vehicle, CRS/vehicle, and CRS/D4 treated mice (from top to bottom).The nuclei were stained with DAPI (blue).Scale bar, 100 µm.dDG: dorsal dentate gyrus; vDG: ventral dentate gyrus; vCA1: ventral cornu ammonis 1; vSub: ventral subiculum; vEC: ventral entorhinal cortex; LS: lateral septum; Amy: amygdala; NAc: nucleus accumbens.F Violin plot of c-Fos + cells in each region after the TST.Ctrl/vehicle, white; CRS/vehicle, red; CRS/D4, blue.The open circle indicates the value from each section.n = 15 sections from 5 mice for all groups.One-way ANOVA followed by LSD post hoc test (LS and NAc).Kruskal-Wallis test with Bonferroni correction (dDG, vDG, vCA1, vSub, EC, and Amy).*p < 0.05, **p < 0.01, ***p < 0.001