Essential role of microglial transforming growth factor-β1 in antidepressant actions of (R)-ketamine and the novel antidepressant TGF-β1

In rodent models of depression, (R)-ketamine has greater potency and longer-lasting antidepressant effects than (S)-ketamine; however, the precise molecular mechanisms underlying the antidepressant actions of (R)-ketamine remain unknown. Using RNA-sequencing analysis, we identified novel molecular targets that contribute to the different antidepressant effects of the two enantiomers. Either (R)-ketamine (10 mg/kg) or (S)-ketamine (10 mg/kg) was administered to susceptible mice after chronic social defeat stress (CSDS). RNA-sequencing analysis of prefrontal cortex (PFC) and subsequent GSEA (gene set enrichment analysis) revealed that transforming growth factor (TGF)-β signaling might contribute to the different antidepressant effects of the two enantiomers. (R)-ketamine, but not (S)-ketamine, ameliorated the reduced expressions of Tgfb1 and its receptors (Tgfbr1 and Tgfbr2) in the PFC and hippocampus of CSDS susceptible mice. Either pharmacological inhibitors (i.e., RepSox and SB431542) or neutralizing antibody of TGF-β1 blocked the antidepressant effects of (R)-ketamine in CSDS susceptible mice. Moreover, depletion of microglia by the colony-stimulating factor 1 receptor (CSF1R) inhibitor PLX3397 blocked the antidepressant effects of (R)-ketamine in CSDS susceptible mice. Similar to (R)-ketamine, the recombinant TGF-β1 elicited rapid and long-lasting antidepressant effects in animal models of depression. Our data implicate a novel microglial TGF-β1-dependent mechanism underlying the antidepressant effects of (R)-ketamine in rodents with depression-like phenotype. Moreover, TGF-β1 and its receptor agonists would likely constitute a novel rapid-acting and sustained antidepressant in humans.


Introduction
In 1990, Trullas and Skolnick 1 demonstrated that Nmethyl-D-aspartate receptor (NMDAR) antagonists such as (+)-MK-801 showed antidepressant-like effects in rodents. In 2000, Berman et al. 2 demonstrated the rapidacting and sustained antidepressant effects of the NMDAR antagonist ketamine in patients with major depressive disorder (MDD). Subsequently, several groups replicated the robust antidepressant effects of ketamine in treatment-resistant patients with either MDD or bipolar disorder [3][4][5][6][7][8][9][10] . Interestingly, ketamine rapidly reduced suicidal thoughts in depressed patients with suicidal ideation within 1 day and for up to 1 week 11,12 . In addition, it is suggested that suicidal thoughts may be related to symptoms of anhedonia independent of other depressive symptoms 13 . Meta-analyses revealed that ketamine has rapid-acting and sustained antidepressant effects and antisuicidal ideation effects in treatment-resistant patients with depression [14][15][16] . Importantly, meta-analyses showed that the effect sizes of ketamine are larger than those of other NMDAR antagonists 14,15 , suggesting that NMDAR blockade is not a sole mechanism of antidepressant action for ketamine. The collective rapid-acting and sustained antidepressant actions of ketamine in depressed patients are serendipitous in the field of psychiatry; [17][18][19] however, the precise molecular and cellular mechanisms underlying antidepressant effects of ketamine remain to be elucidated [20][21][22][23][24][25] . Off-label use of ketamine is popular in the United States (US), although the adverse side-effects (i.e., psychotomimetic effects, dissociation, and abuse liability) of ketamine remain to be resolved 26,27 .
Ketamine (Ki = 0.53 μM for NMDAR), also known as (R, S)-ketamine, is a racemic mixture that contains equal amounts of (R)-ketamine (or arketamine) (Ki = 1.4 μM for NMDAR) and (S)-ketamine (or esketamine) (Ki = 0.30 μM for NMDAR) 24 . Preclinical data have shown that (R)-ketamine displays greater potency and longer -lasting antidepressant effects than (S)-ketamine in rodent models of depression [28][29][30][31][32][33][34] , suggesting that NMDARs do not play a major role in the robust antidepressant effects of ketamine 24 . Importantly, in both rodents and monkey, the sideeffects of (R)-ketamine were lower than were those of (R,S)ketamine and (S)-ketamine 29,[35][36][37][38] . In addition, in humans, the incidence of psychotomimetic side-effects of (S)-ketamine (0.45 mg/kg) was higher than that of (R)-ketamine (1.8 mg/kg), although the dose of (S)-ketamine was lower than was that of (R)-ketamine 39 . Though (S)-ketamine produced psychotic reactions, including depersonalization and hallucinations, the same dosage of (R)-ketamine did not induce psychotic symptoms in the healthy subjects, and most of them experienced a state of relaxation 40 . These results indicate that (S)-ketamine contributes to the acute side-effects of ketamine, whereas (R)-ketamine may not be associated with these side-effects 22,24 . On 5 March 2019, the US Food & Drug Administration approved (S)-ketamine nasal spray for treatment-resistant depressed patients. Due to the risk of serious adverse effects, (S)-ketamine nasal spray can be obtained only through a restricted distribution system under the Risk Evaluation and Mitigation Strategy. A clinical trial of (R)-ketamine in humans is underway 24 . Meanwhile, little is known about the precise molecular mechanisms underlying the different antidepressant effects of the two enantiomers 24,25,41,42 .
The aim of this study was to identify the novel molecular mechanisms underlying the antidepressant effects of (R)-ketamine in animal models of depression. First, we conducted RNA-sequencing analysis of the prefrontal cortex (PFC) of chronic social defeat stress (CSDS) susceptible mice treated with either (R)-ketamine or (S)ketamine, as PFC contributes to the antidepressant actions of ketamine and its enantiomers 29,43,44 . Second, we studied the effects of pharmacological inhibitors and a neutralizing antibody of the novel target in the antidepressant effects of (R)-ketamine. Finally, we investigated whether the novel molecule (i.e., TGF-β) has rapid-acting and sustained antidepressant effects in rodent models of depression.

Animals
Male adult C57BL/6 mice, aged 8 weeks (body weight 20-25 g, Japan SLC, Inc., Hamamatsu, Japan), male CD1 mice, aged 14 weeks (body weight 40-45 g, Japan SLC, Inc., Hamamatsu, Japan) were used in the experiments. Male Sprague-Dawley rats, aged 7 weeks (body weight 200-230 g, Charles-River Japan, Co., Tokyo, Japan) were used for learned helplessness (LH) model. No blinding for animal experiments was done. Animals were housed under controlled temperature and 12 h light/dark cycles (lights on between 07:00-19:00), with ad libitum food and water. The study was approved by the Chiba University Institutional Animal Care and Use Committee.

CSDS model and LPS-induced model
The procedure of CSDS was performed as previously reported 29,[32][33][34]47 . Detailed methods were shown in the supplemental information.

Gene expression analysis by quantitative real-time PCR
Control mice and CSDS susceptible mice were sacrificed 3 days after intraperitoneal (i.p.) administration of saline (10 ml/kg), (R)-ketamine (10 mg/kg), or (S)-ketamine (10 mg/kg). The PFC and hippocampus were quickly dissected on ice from whole brain since these brain regions play a key role in antidepressant effects of (R)-ketamine 44 . Detailed methods were shown in the supplemental information.
To examine the effects of microglia depletion, PLX3397 (100 μM, 2 μl, i.c.v.) or vehicle (10% DMSO and 90% SBEβ-CD) was administered to mice under isoflurane anesthesia. PFC was collected 24 h after injection. Right PFC and left PFC were used for FACS analysis and Western blot of Iba1, respectively.
For intranasal administration, saline (15 μl) or TGF-β1 (1.5 μg, 15 μl) was administered to mice 23 hrs after LPS administration, as previously reported 38 . Mice were restrained by hand, and saline or TGF-β1 was administered intranasally into awake mice using Eppendorf micropipette (Eppendorf Japan, Tokyo, Japan). The locomotion and FST were performed 1 and 3 h after injection, respectively.

Learned helplessness (LH) model
Rat LH paradigm was performed as previously reported 44,54 . Detailed methods were shown in the supplemental information.

Western blot analysis of Iba1
Western blot analysis was performed as reported previously 29,34,51,52 . Detailed methods were shown in the supplemental information.

FACS analysis
Mouse PFC tissues were mashed and passed through a 70 μm mesh to prepare single cell suspension then subjected for FACS analysis. Cells were stained with monoclonal antibodies against cell surface antigens at 4°C for 30 min, then washed with PBS. In indicated cells, cells were fixed and permeabilized using FoxP3 staining buffer set (Invitrogen) according to the manufacturer instruction. Then intracellular antigens were stained with indicated antibodies at room temperature for 30 min. The following antibodies were used for staining; anti TMEM119-PE (Abcam, Cambridge, UK), allophycocyanin conjugated anti CD11b (BD Bioscience, Franklin Lakes, NJ), anti Iba1-FITC (Abcam), anti TGFβ-allophycocyanin (BioLegend, San Diego, CA). The stained cells were analyzed using FACSCantII and FlowJo software (BD).

Statistical analysis
The data show as the mean ± standard error of the mean (S.E.M.). Analysis was performed using PASW Statistics 20 (formerly SPSS Statistics; SPSS). The data were analyzed using Student t-test or the one-way analysis of variance (ANOVA), followed by post hoc Tukey test. The P-values < 0.05 were considered statistically significant.

RNA-sequencing analysis of PFC samples
To identify the novel molecular targets for the antidepressant effects of (R)-ketamine, we collected PFC samples 3 days after either (R)-ketamine (10 mg/kg) or (S)-ketamine (10 mg/kg) were administered to CSDS susceptible mice. We performed RNA-sequencing analysis of PFC samples from animals treated with either (R)ketamine or (S)-ketamine (Fig. 1a). GSEA revealed that TGF-β signaling might be involved in the differential antidepressant effects of the two enantiomers (Fig. 1b). We found reduced expression of Tgfb1 and its receptors (Tgfbr1 and Tgfbr2) in the PFC and hippocampus from CSDS susceptible mice (Fig. 1c-f and Fig. S1). Conversely, the expression of Tgfb2 in the PFC and the hippocampus did not differ in the four groups (Fig. 1c-f and Fig. S1). Interestingly, (R)-ketamine (10 mg/kg), but not (S)-ketamine (10 mg/kg), significantly ameliorated the reduced expression of these genes (Fig. 1c-f and Fig. S1).

Role of microglial TGF-β1
TGF-β1 is constitutively expressed in microglia into adulthood 55 . An earlier study demonstrated that TGF-β1 was necessary for the in vitro development of microglia and that microglia were absent in the brain of TGF-β1deficient mice 56 , suggesting that TGF-β1 plays a key role in microglia. Microglia rely on cytokine signaling, such as activation of CSF1R and TGF-β1, for their survival 57 . In situ hybridization with cell-type marker immunostaining revealed high expression of Tgfb1 and its receptors (Tgfbr1 and Tgfbr2) in microglia, but not in astrocytes, in mouse brain PFC (Fig. 3).
To examine whether microglia TGF-β1 contributes to the antidepressant effects of (R)-ketamine, we studied the impact of microglial depletion on the antidepressant effects of (R)-ketamine. Preliminary experimentation revealed that i.c.v. injection of PLX3397, a potent CSF1R inhibitor, reduced the Iba1 protein in the mouse PFC (Fig.  S2). In this study, we used the time (24 h) of PLX3397 (100 μM, 2 μl, i.c.v.). Using FACS analysis, we analyzed the expression of both Iba1 and TGF-β1 in TMEM119 + CD11b + microglia in the PFC. Pretreatment with PLX3397 significantly reduced the expression of both TGF-β1 and Iba1 in TMEM119 + CD11b + microglia (Fig. 4a-c). Furthermore, Western blot analysis revealed that PLX3397 injection reduced Iba1 protein in the PFC (Fig. 4d). These findings indicate partial depletion of microglia by PLX3397 in the PFC.
Next, we studied the impact of PLX3397 on the antidepressant effects of (R)-ketamine in CSDS susceptible mice (Fig. 5a). There were no changes in locomotion among the five groups (Fig. 5b). Findings from the TST and the forced swim test (FST), showed that PLX3397 significantly blocked the antidepressant effects of (R)-ketamine for increased immobility time of both TST and FST (Fig. 5c, d). In the SPT, PLX3397 significantly blocked the effects of (R)ketamine for reduced sucrose preference in CSDS susceptible mice (Fig. 5e). Collectively, partial depletion of microglia by PLX3397 significantly blocked the antidepressant effects of (R)-ketamine in CSDS susceptible mice (Fig. 5). These findings indicate that microglia-expressing molecules, including TGF-β1 and its receptors, contribute to the antidepressant effects of (R)-ketamine in a CSDS model.

Antidepressant effects of TGF-β1 in rodent models of depression
Finally, we studied whether mouse recombinant TGF-β1 has antidepressant effects in three animal models of depression. First, we studied the effects of TGF-β1 and TGF-β2 in the CSDS model (Fig. 6a). There were no changes in locomotion in the four groups (Fig. 6b, h). A single i.c.v. injection of (R)-ketamine (1 mg/ml, 2 μl) produced rapid and sustained antidepressant effects in CSDS susceptible mice, consistent with the previous report 58 . Similar to (R)-ketamine, i.c.v. infusion of TGF-β1 (10 ng/ml, 2 μl) significantly the increased Fig. 3 In situ hybridization and immunohistochemistry. a Representative image of Tgfb1 mRNA (purple) and Iba1 protein (brown, marker for microglia) or S100b protein (brown, marker for astrocyte). b Representative image of Tgfbr1 mRNA. c Representative image of Tgfbr2 mRNA. Tgfb1 and its receptors (Tgfbr1 and Tgfbr2) are co-localized with microglia, but not astrocytes. Scale bar = 100 μm. immobility time of both TST and FST in CSDS susceptible mice (Fig. 6c, d). In the SPT, i.c.v. infusion of TGF-β1 significantly the reduced sucrose preference in CSDS susceptible mice ( Fig. 6e-g). Interestingly, we detected the beneficial effects of TGF-β1 seven days after a single injection (Fig. 6g), indicating long-lasting antidepressant effects of TGF-β1. Conversely, TGF-β2 (10 ng/ml, 2 μl) did not produce antidepressant effects in CSDS susceptible mice, though (R)-ketamine (1 mg/ ml, 2 μl) produced rapid and sustained antidepressant effects in the same model (Fig. 6h-l).
Moreover, a single i.c.v. infusion of TGF-β1 (10 ng/ml, 2 μl) significantly attenuated the increased immobility time of FST in LPS (0.5 mg/kg)-treated mice (Fig. 7a-c). In addition, a single intranasal administration of TGF-β1 (1.5 μg, 15 μl) significantly attenuated the increased immobility time of FST in LPS-treated mice (Fig. 7d-f). In a rat LH model, bilateral i. c.v. infusion of TGF-β1 (250 ng/side) significantly reduced the failure number and latency of LH rats 4 days after i.c.v. injection (Fig. 7g-i). These findings indicate that recombinant TGF-β1 has ketamine-like robust antidepressant effects in rodent models of depression.

Discussion
The main findings of this study are as follows: First, RNAsequencing and GSEA revealed the role of TGF-β signaling in the beneficial antidepressant effects of (R)-ketamine compared with (S)-ketamine. RT-PCR revealed reduced expression of Tgfb1 and its receptors (Tgfbr1 and Tgfbr2) in the PFC and the hippocampus from CSDS susceptible mice. Furthermore, (R)-ketamine, but not (S)-ketamine, attenuated the reduced expression of these genes in the PFC and the hippocampus of CSDS susceptible mice. Second, pharmacological inhibitors and neutralizing antibody of TGF-β1 blocked the antidepressant effects of (R)-ketamine in CSDS susceptible mice, indicating a role of TGF-β1 signaling in the antidepressant effects of (R)-ketamine. Third, partial depletion of microglia by PLX3397 blocked antidepressant effects of (R)-ketamine in CSDS susceptible mice, indicating a role of microglia in the antidepressant Red histograms indicate control group and blue histograms indicate PLX3397 treated group. c The fluorescence intensity of both Iba1 and TGF-β1 in TMEM119 + CD11b + microglia in the PFC of PLX3397 treated mice was significantly (Iba1: P = 0.0186, TGF-β1: P = 0.0002) lower than that of control mice. d Western blot analysis of Iba1 in the PFC samples of control mice and PLX3397 treated mice. Representative bands of Western blot analysis. The expression of Iba1 in the PFC of PLX3397 treated mice was significantly (P = 0.0023) lower than that of control mice. Data are shown as mean ± SEM. (control group: n = 10, PLX group n = 9). *P < 0.05, **P < 0.01, ***P < 0.001. effects of (R)-ketamine. Lastly, recombinant TGF-β1 elicited rapid-acting and long-lasting antidepressant effects in CSDS, LPS, and LH models of depression. Overall, it appears likely that (R)-ketamine can exert antidepressant effects by normalizing microglial TGF-β1 signaling in the PFC and the hippocampus of CSDS susceptible mice. Furthermore, TGF-β1 has ketamine-like antidepressant effects in rodent models.
Microglia are the only cell type that express CSF1R. CSF1R knockout mice are devoid of microglia 59 . Moreover, it has been reported that repeated treatment with CSF1R inhibitors, such as PLX3397, cause a dramatic reduction in the number of microglia within the adult brain [48][49][50] .
Interestingly, microglia are absent in the brains of central nervous system TGF-β1 knockout mice 56 . Thus, microglia in the adult brain are physiologically dependent upon CSF1R and TGF-β1 signaling 57 . In this study, a single i.c.v. injection of PLX3397 produced significant reduction of Iba1 and TGF-β1 in the PFC, suggesting partial depletion of microglia in the PFC. Interestingly, pretreatment of PLX3397 significantly blocked the antidepressant effects of (R)-ketamine in CSDS susceptible mice. Overall, it appears likely that microglial TGF-β1 in the PFC might contribute to the antidepressant effects of (R)-ketamine.
In this study, i.c.v. infusion of TGF-β1 produced rapidacting and long-lasting antidepressant effects in a CSDS Fig. 5 Effects of PLX3397 on antidepressant effects of (R)-ketamine in a CSDS model. a Chronic social defeat stress (CSDS) was performed from day 1 to day 10 for 10 days. Social interaction test was performed on day 11. On day 12, vehicle or PLX3397 was administered i.c.v. to CSDS susceptible mice. On day 13, saline or (R)-ketamine (10 mg/kg) was administered i.p. 24 h after injection of PLX3397. Locomotion and FST were performed 1 and 3 h after injection, respectively. FST and SPT were performed 1 and 2 days after injection, respectively. b Locomotion (1 h, one-way ANOVA, model, an LPS-induced model, and an LH model. Notably, we detected the antidepressant effects of TGF-β1 in a CSDS model and an LH model 7 days and 4 days after a single dose, respectively. Collectively, the antidepressant effects of TGF-β1 in these models are similar to those of (R)-ketamine, suggesting that TGF-β1 has (R)-ketamine-like longlasting antidepressant effects. Taylor et al 60 . showed that a single i.c.v. injection of TGF-β1 4 h after intracerebral hemorrhage (ICH) produced complete recovery of motor function at 24 h, and that this recovery persisted for at least one week. Furthermore, i.c.v. injection of TGF-β1 alleviated N-methyl-4-phenylpyridinium ion (MPP + )-induced microglial inflammatory response and dopaminergic neuronal loss in the substantia nigra, indicating that TGF-β1 plays a role in the pathology of Parkinson's disease (PD). Collectively, it is possible that TGF-β1 can produce rapid and long-lasting beneficial effects in several models, such as depression, ICH, and PD.
Notably, intranasal administration of TGF-β1 has rapidacting antidepressant effects in LPS-treated mice. A previous study showed that intranasal administration of TGF-β1 ameliorated neurodegeneration in the mouse brain after β-amyloid 1-42 injection 44 . It has also been reported that TGF-β1 administered intranasally entered several brain regions, such as the PFC and the hippocampus, of control adult mice, whereas no increase was observed in the blood and peripheral organs 61 , indicating good permeability of the blood brain barrier for TGF-β1. It is also reported that CSDS alters blood brain barrier integrity through loss of tight junction protein Cldn5 62 . In addition, TGF-β1 might be free of the psychotomimetic side-effects of ketamine and its potential for abuse in humans, as TGF-β1 does not interact with NMDAR in the brain. Therefore, it is likely that intranasal administration of TGF-β1 would be a novel potential therapeutic approach for depression. This study has some limitations. In this study, we used the CSF1R inhibitor to delete microglia in the brain although the partial depletion of microglia was detected. It is of great interest to investigate the role of microglia in the antidepressant effects of (R)-ketamine using CSF1R knockout mice since CSF1R knockout mice are devoid of microglia 59 . Furthermore, it is also of interest to investigate the role of microglial TGF-β1 in the antidepressant Fig. 6 Effects of recombinant TGF-β1 and TGF-β2 in a CSDS model. a Chronic social defeat stress (CSDS) was performed from day 1 to day 10 for 10 days. Social interaction test was performed on day 11. On day 12, vehicle or TGF-β1 (or TGF-β2) was administered i.c.v. to CSDS susceptible mice. Locomotion and TST were performed 1 and 3 h after injection, respectively. FST was performed 1 day after injection. SPT was performed 2, 4, and 7 days after injection. b Locomotion (1 h, one-way ANOVA, F 3,20 = 0.122, P = 0.946). c TST (3 h, one-way ANOVA, F 3,20 = 2.352, P = 0.041). d FST (1 day, one-way ANOVA, F 3,20 = 3.650, P = 0.030). e SPT (2 day, one-way ANOVA, F 3,20 = 3.410, P = 0.037). f SPT (4 day, one-way ANOVA, F 3,20 = 8.140, P = 0.001). g SPT (7 day, one-way ANOVA, F 3,20 = 6.278, P = 0.004). h Locomotion (1 h, one-way ANOVA, F 3,20 = 0.171, P = 0.975). i TST (3 h, one-way ANOVA, F 3,20 = 10.093, P < 0.001). j FST (1 day, one-way ANOVA, F 3,20 = 16.353, P < 0.001). k SPT (2 day, one-way ANOVA, F 3,20 = 4.750, P = 0.012). l SPT (4 day, one-way ANOVA, F 3,20 = 5.404, P = 0.007). Data are shown as mean ± SEM. (n = 6). *P < 0.05, **P < 0.01. ANOVA analysis of variance, FST forced swimming test, N.S. not significant, R-KT (R)-ketamine, SPT sucrose preference test, TST tail suspension test. effects of (R)-ketamine using TGF-β1 knockout mice since microglia were absent in the brain of TGF-β1 knockout mice 56 .
In conclusion, this study shows that TGF-β1 in the microglia might contribute to the antidepressant effects of (R)-ketamine in animal models of depression. Furthermore, similar to (R)-ketamine, TGF-β1 seems to rapidacting and long-lasting antidepressant effects. Therefore, it is likely that TGF-β1 would be a new rapid-acting and sustained antidepressant. Fig. 7 Effects of recombinant TGF-β1 in LPS model and LH model. a Saline or LPS (0.5 mg/kg) was administered i.p. to mice. Saline or TGF-β1 was administered i.c.v. to LPS-treated mice 23 h after LPS injection. Locomotion and FST were performed 1 and 3 h after injection, respectively. b Locomotion (1 h, one-way ANOVA, F 2,39 = 0.122, P = 0.122). c FST (3 h, one-way ANOVA, F 2,39 = 3.124, P = 0.045). Data are shown as mean ± SEM. (n = 14). *P < 0.05. d Saline or LPS (0.5 mg/kg) was administered i.p. to mice. Saline or TGF-β1 was administered intranasally to LPS-treated mice 23 h after LPS injection. Locomotion and FST were performed 1 and 3 h after injection, respectively. e Locomotion (1 h, one-way ANOVA, F 2,27 = 0.255, P = 0.777). f FST (3 h, one-way ANOVA, F 2,27 = 5.180, P = 0.013). Data are shown as mean ± SEM. (n = 10). *P < 0.05, **P < 0.01. g Rats received inescapable electric stress shock (IES) treatments on 2 days (day 1 and day 2), passed a post-shock test (PS) on day 3 to select learned helplessness (LH) rats with depression-like phenotype. On day 4, vehicle or TGF-β1 was administered i.c.v. into LH rats. On day 8 (4 days after i.c.v. injection), conditioned avoidance (CA) tests to study the antidepressant effect was performed. h The failure number of TGF-β1 treated LH rats was significantly (P = 0.0259) lower than that of vehicle treated LH rats. i The escape latency of TGF-β1 treated LH rats was significantly (P = 0.0281) lower than that of vehicle treated LH rats. Data are shown as mean ± SEM. (vehicle: n = 5, TGF-β1: n = 6). *P < 0.05. FST forced swimming test, N.S. not significant.