Tofacitinib prevents depressive-like behaviors through decreased hippocampal microgliosis and increased BDNF levels in both LPS-induced and CSDS-induced mice

Abstract

Depressive disorders are a global mental health challenge that is closely linked to inflammation, especially in the post-COVID-19 era. The JAK-STAT pathway, which is primarily associated with inflammatory responses, is not fully characterized in the context of depressive disorders. Recently, a phase 3 retrospective cohort analysis heightened that the marketed JAK inhibitor tofacitinib is beyond immune diseases and has potential for preventing mood disorders. Inspired by these clinical facts, we investigated the role of the JAK-STAT signaling pathway in depression and comprehensively assessed the antidepressant effect of tofacitinib. We found that aberrant activation of the JAK-STAT pathway is highly conserved in the hippocampus of classical depressive mouse models: LPS-induced and chronic social defeat stress (CSDS)-induced depressive mice. Mechanistically, the JAK-STAT pathway mediates proinflammatory cytokine production and microgliosis, leading to synaptic defects in the hippocampus of both depressive models. Remarkably, the JAK inhibitor tofacitinib effectively reverses these phenomena, contributing to its antidepressant effect. These findings indicate that the JAK/STAT pathway could be implicated in depressive disorders, and suggest that the JAK inhibitor tofacitinib has a potential translational implication for preventing mood disorders far beyond its current indications.

Introduction

Major depressive disorder (MDD) is the most common mental disorder, with a prevalence of 16% in modern society [1]. It is characterized by a high disability rate, high recurrence rate, and high suicide rate [1, 2]. The core features of MDD include persistent sadness, anhedonia, loss of motivation and passion in life, severely diminished quality of life, and even life threats [3]. According to World Health Organization data, in 2022, the global prevalence of anxiety and depressive disorders increased by 25% during the first year of the COVID-19 pandemic [4]. In addition, up to 50% of patients with MDD do not respond to current first-line antidepressant medications and do not achieve remission [5]. Based on these refractory features, it is necessary to prioritize the identification and prevention of depression globally.

Clinical experiments have suggested a correlation between inflammation and depression [6, 7]. A significant increase in neuroinflammatory levels was observed in clinically MDD patients [8]. Activated microglia have also been detected in brain samples from MDD patients who have committed suicide [9]. Recent studies have proposed that increased levels of inflammation are implicated in the onset of depression. Clinical trials have shown that subjects vaccinated with Salmonella typhi exhibit depressive symptoms as IL-6 levels increase [10]. Another study indicated that contracting autoimmune diseases increases the risk of developing mood disorders [11]. Therefore, targeting inflammation in the brain could prevent the onset of depression.

The JAK-STAT signaling pathway is responsive to many proinflammatory cytokines and plays a crucial role in initiating innate immunity and coordinating adaptive immunity [12]. Once the JAK-STAT signaling pathway is activated, it can further increase the transcription of proinflammatory cytokines and aggravate the level of inflammation [13]. However, current research on JAK-STAT signaling has focused mainly on peripheral diseases such as rheumatoid arthritis, and less research has been conducted on emotional disorders [14]. Recent studies based on protein interaction network analysis have confirmed that upregulation of the JAK-STAT signaling pathway may contribute to the onset of depression [15]. A significant increase in JAK3 activity was reported to induce the reduction in neurogenesis in the hippocampus in a glucocorticoid-induced stress mouse model [16]. Clinical observations have also suggested that the efficacy of antidepressant treatment may be related to the regulation of the JAK-STAT signaling pathway. For example, the tricyclic antidepressant amitriptyline can reduce JAK3 phosphorylation in the hippocampus, improving psychological health and quality of life in patients with inflammatory bowel disease [17].

These findings suggest a close association between the JAK-STAT signaling pathway and depression, making it a potential target for depression prevention. Therefore, to investigate whether pharmacological inhibition of JAK-STAT signaling can prevent depression, we chose the marketed drug tofacitinib. The pan-JAK inhibitor tofacitinib, developed by Pfizer, showed a greater ability to cross the blood‒brain barrier than other JAK inhibitors [18]. Preclinical studies have confirmed that tofacitinib can effectively inhibit the activity of JAK1 and JAK3, which are closely associated with the signal transduction of neuroinflammatory cytokines [19]. Additionally, a phase III retrospective analysis revealed that tofacitinib was found to significantly improve depression in patients with rheumatoid arthritis compared with other drug-treated groups [20]. These clinical findings suggested that tofacitinib has potential anti-depressant effects, which requires systematic evaluation.

In this study, we investigated the role of JAK-STAT pathway in depression and comprehensively assessed the anti-depressant effect of tofacitinib, using two classic mouse models of depression [21, 22]. The hippocampus we studied is susceptible to stress conditions and traditional conditions associated with the onset of depression [23]. In this region, we mechanistically revealed the activation of the JAK-STAT signaling pathway and subsequent changes in microgliosis and synaptic functions in both models of depression. From a translational perspective, we utilized the marketed drug tofacitinib to evaluate its protective effects on depressive-like behaviors in both mouse models. Furthermore, we revealed that the anti-depressant effects of tofacitinib are attributed to the modulation of microgliosis and synaptic functions in the hippocampus, which provides potentially translational implications.

Results

JAK/STAT pathway is activated in the hippocampus of LPS-induced depressive mice

Although the activation of JAK/STAT signaling has been reported the cause of autoimmune encephalitis, its role in depression is incompletely characterized [24]. To this end, we initially established a depressive-like mouse model by intraperitoneal injection of LPS (Fig. 1a). This model was proposed to exhibit neuroinflammation and hence mimic the depressive symptoms of humans [25], which might be linked to the inflammatory reaction of the JAK/STAT pathway. In the post-injection, the body weight of the LPS-induced mice was significantly lower than that of the control mice, indicating the occurrence of sickness behaviors (Fig. 1b). Depressive mental status is among the most common sickness behaviors; thus, we subjected these mice to the forced swim test (FST) and tail suspension test (TST) to assess their depressive-like behaviors (Fig. 1c, d). We found that immobility time in the FST and TST was significantly increased in LPS-induced mice, compared to control mice (Fig. 1c, d). Moreover, we employed the open field test (OFT), which is a frequently used method to estimate locomotor activity in rodents [26]. In the OFT, no significant difference in total distance traveled was observed between the different groups (Fig. 1e). These behavioral results demonstrated that LPS challenge successfully induced depressive-like behavior but did not alter locomotor ability in mice. Given the inflammatory reaction of the JAK/STAT pathway, we next wondered whether this pathway is involved in the development of depressive-like behaviors in LPS-induced mice. To this end, we assessed the phosphorylation levels of JAKs and STATs in the hippocampus, focusing on the JAK1 and JAK3 targeted by tofacitinib [19]. We found that the phosphorylation level of JAK3 was significantly increased, and the phosphorylation level of JAK1, which forms a dimer with JAK3, also showed a tendency to increase (Fig. 1f–h). The phosphorylation level of STAT1 and STAT3, which is the downstream of JAK3, was significantly increased in the hippocampus of LPS-induced mice (Fig. 1i, j). Interestingly, no significant difference in the phosphorylation of STAT5 were observed, which may be attributed to the low abundance of STAT5 in the brain [27] (Fig. 1k). These results indicated that activation of the JAK/STAT pathway is closely associated with the development of depressive-like behavior in LPS-induced mice.

Fig. 1. The JAK/STAT pathway is involved in the progression of depressive-like behaviors in LPS-induced mice.

a Experimental timeline is shown for in vivo LPS-induced depression mice model. b Relative body weights differences. (n = 9). c Immobility time in the forced swimming test, the immobility time in the last 4 min was analyzed (n = 9). d Immobility time in the tail suspension test, the immobility time in the last 4 min was analyzed (n = 9). e Total distance traveled in the open field test. (n = 9). f Representative blots showing JAKs and STATs expression (n = 3). g–k Representative bar graph of JAKs and STATs expression analysis (n = 3). Data are expressed as mean ± SEM, *P < 0.05, **P < 0.01.

The JAK inhibitor tofacitinib alleviates depressive-like behavior in LPS-induced mice

Based on these intriguing findings, we wondered whether pharmacological inhibition of the JAK/STAT pathway could improve depressive-like behavior in LPS-induced mice. Tofacitinib, an FDA-approved JAK inhibitor, has the ability to cross the blood‒brain barrier and has shown potential for preventing mood disorders in recent clinical reports [20, 28]; thus, we chose to use tofacitinib for investigation. The choice of tofacitinib dose was based on its FDA approval databases that 100 mg/kg higher dose showed noticeably negative effect on central nervous system in mice. Thus, we chose the dose of 10 mg/kg, 30 mg/kg as reported previously [29], and explored the dose of 75 mg/kg owing to the low cerebral bioavailability of most drugs [30]. To exclude drug’s side-effects on emotion itself, the naïve mice were treated with tofacitinib (10, 30, and 75 mg/kg) or CMC-Na as a control by gavage for 1 week (Supplementary Fig. 1a). We found that different doses of tofacitinib insignificantly change body weight, locomotor activity and emotional behaviors in the naïve mice (Supplementary Fig. 1b–e). These results indicated that 10, 30, and 75 mg/kg tofacitinib is well-tolerated, while unaffecting emotional status in normal conditions. To assess its antidepressant effect, the LPS-induced depressive mice were intragastrically treated with tofacitinib (10, 30, and 75 mg/kg) or CMC-Na as a control once a day for 1 week (Fig. 2a). We found that 10 mg/kg and 30 mg/kg tofacitinib showed a protective effect on LPS-induced weight loss in a dose-dependent manner. However, the effect of 75 mg/kg tofacitinib on body weight was limited, suggesting that the high dose of tofacitinib in the inflammatory state has some side-effects (Fig. 2b). Moreover, we found that 10 mg/kg tofacitinib showed a tendency for improvement in depression, and both 30 mg/kg and 75 mg/kg tofacitinib significantly attenuated depressive-like behaviors in LPS-treated mice (Fig. 2c, d). In addition, tofacitinib treatment did not alter locomotor ability in the OFT in the different groups (Fig. 2e). Given that 30 mg/kg tofacitinib had a greater protective effect on sickness behavior (as indicated by body weight) and similar improvements in depressive status compared to 10 and 75 mg/kg tofacitinib, we chose a 30 mg/kg dose for further investigations. Consistent with previous findings (Fig. 1f, g), the phosphorylation of JAK3, STAT1, and STAT3 was significantly increased in the hippocampus of LPS-induced mice, but this phenomenon was reversed by 30 mg/kg tofacitinib (Fig. 2f–k). Overall, our findings revealed that the JAK inhibitor tofacitinib could effectively mitigate aberrant activation of the JAK/STAT pathway in the hippocampus, leading to behavioral improvements in LPS-induced mice.

Fig. 2. The JAK inhibitor tofacitinib effectively mitigates depressive-like behavior in LPS-induced mice.

a Schematic timeline of LPS, tofacitinib treatment and behavioral tests. b Relative body weights differences. (n = 7–17). c Immobility time in the forced swimming test, the immobility time in the last 4 min was analyzed (n = 7–17). d Immobility time in the tail suspension test, the immobility time in the last 4 min was analyzed (n = 7–17). e Total distance traveled in the open field test. (n = 7–17). f Representative blots showing JAKs and STATs expression (n = 3). g–k Representative bar graph of JAKs and STATs expression analysis (n = 3). Data are expressed as mean ± SEM, *P < 0.05, **P < 0.01.

The antidepressant effect of tofacitinib is attributed to decreased hippocampal microgliosis and increased BDNF levels in LPS-induced mice

The JAK/STAT pathway is involved in the production of neuroinflammatory cytokines, such as C-C motif ligand 2 (CCL2), interleukin-1β (IL-1β) and tumor necrosis factor-α (TNF-α), which lead to depression [31–33]. Based on these studies, we next evaluated the levels of JAK/STAT-mediated cytokines in LPS-induced mice. As expected, the transcription of cytokines, including CCL2, IL-1β and TNF-α, was significantly upregulated in the hippocampus, where the number of microglia was also increased in LPS-induced mice (Fig. 3a–c). Morphological analysis of microglia also highlighted an increase in soma area, while the length of microglial processes, which is neuroinflammatory phenotypes of microglia (Supplementary Fig. 2a–c). These findings were consistent with the notion that cytokines are mediators of the status of microglia and the neuroimmune system in depression [33]. Aberrant activation of JAK/STAT has also been reported to participate in inflammation-driven synaptic defects in autoimmune encephalitis [24], so we further assessed BDNF levels (a sign of synaptic defects) in LPS-induced mice. We found that LPS challenge also significantly decreased BDNF mRNA and protein levels in the hippocampus (Fig. 3d–f). Interestingly, when LPS-induced mice were intragastrically injected with 30 mg/kg tofacitinib, JAK/STAT-mediated cytokine transcription and microgliosis were significantly decreased in the hippocampus (Fig. 3g–i and Supplementary Fig. 3a–c). Moreover, these tofacitinib-treated mice were also more resilient to the LPS-induced decrease in BDNF levels in hippocampus (Fig. 3j–l). Overall, tofacitinib has multiple functions, including preventing hippocampal microgliosis and increasing the level of BDNF, which contributes to its antidepressant effect on the inflammatory context.

Fig. 3. Decreased hippocampal microgliosis and increased BDNF levels contribute to the antidepressant effect of tofacitinib in LPS-induced mice.

a Representative immunofluorescence of Iba-1 in the hippocampus. b Column graphs representing Iba-1 expression (n = 3). c Analysis of CCL2, IL-1β, TNF-α mRNA levels in the hippocampus (n = 6). d Analysis of BDNF mRNA levels in the hippocampus (n = 6). e Representative blots showing BDNF expression (n = 3). f Representative bar graph of BDNF expression analysis (n = 3). g Representative immunofluorescence of Iba-1 in the hippocampus. h Column graphs representing Iba-1 expression (n = 3). i Analysis of CCL2, IL-1β, and TNF-α mRNA levels in the hippocampus (n = 6). j Representative bar graph showing BDNF expression (n = 3). k Representative blots of BDNF expression analysis (n = 3). l Representative bar graph of BDNF expression analysis (n = 3). Data are expressed as mean ± SEM, *P < 0.05, **P < 0.01.

Aberrant activation of the JAK/STAT pathway is highly conserved in CSDS-induced depressive-like mice

To evaluate whether the activation of the JAK/STAT pathway had a similar effect on chronic stress-induced depressive mice, a classical murine CSDS model was generated [22]. These mice were subjected to the CSDS procedure for 10 days, after which their body weight and depressive status were measured (Fig. 4a). The progression of CSDS decreased the social ratio and internation time of the mice, as indicated by the results of the social interaction test (SIT) (Fig. 4b). Unlike those of the LPS-induced mice, the body weights of the CSDS-induced mice did not significantly differ from those of the control mice (Fig. 4c), suggesting that both depressive models complement each other to interrogate the JAK/STAT pathway. To examine their depressive status, these CSDS-induced mice were subjected to the FST and TST. We found that CSDS-induced mice spent more time immobile than control mice in both behavioral tests, indicating the occurrence of depressive-like behaviors (Fig. 4d, e). Moreover, the OFT results showed a decrease in locomotor activity in CSDS-induced mice (Fig. 3f), which might be attributed to chronic exposure to stress, as reported previously [34, 35]. To investigate the connection between the JAK/STAT pathway and depression under chronic stress conditions, we further analyzed the phosphorylation of JAK and STAT in the hippocampus. We observed a marked increase in the phosphorylation of JAK3, STAT1, and STAT3, while the phosphorylation of JAK1 and STAT5 remained unchanged in CSDS-induced mice (Fig. 4g–m). Moreover, the hippocampus of these mice exhibited microgliosis and increased transcription of neuroinflammatory cytokines (Fig. 4n–p and Supplementary Fig. 4a–c). In addition, CSDS procedure also led to a significant reduction in BDNF mRNA and protein levels in the hippocampus (Fig. 4q–s). Interestingly, most of the observations in CSDS-induced mice were similar to those in LPS-induced mice. These results demonstrated that aberrant activation of JAK/STAT signaling in the hippocampus was highly conserved in both LPS-induced and CSDS-induced depressive mice.

Fig. 4. Dysregulation of the JAK/STAT pathway in the hippocampus is highly conserved in CSDS-induced depressive-like mice.

a Experimental timeline is shown for in vivo CSDS-induced depression mice model. b Social index and time in interaction zone in the social interaction test (n = 9–10). c Relative body weight differences (n = 9–10). d Immobility time in the forced swimming test, the immobility time in the last 4 min was analyzed (n = 9–10). e Immobility time in the tail suspension test, the immobility time in the last 4 min was analyzed (n = 9–10). f Total distance traveled in the open field test. (n = 9–10). g representative blots showing JAKs and STATs expression (n = 3). h–l Representative bar graph of JAKs and STATs expression analysis (n = 3). m Representative immunofluorescence of Iba-1 in the hippocampus. n Column graphs representing Iba-1 expression (n = 3). o Analysis of CCL2, IL-1β, TNF-α mRNA levels in the hippocampus (n = 6). p Analysis of BDNF mRNA levels in the hippocampus (n = 6). q Representative blots showing BDNF expression (n = 3). r Representative bar graph of BDNF expression analysis (n = 3). Data are expressed as mean ± SEM, *P < 0.05, **P < 0.01.

The JAK inhibitor tofacitinib has a similar effect on improving depressive-like behaviors in CSDS-induced mice

Based on the highly conserved features of the JAK/STAT pathway, we further wondered whether chronic stress-induced depression could also be prevented by the JAK inhibitor tofacitinib. With this aim, mice were subjected to the CSDS procedure for 10 days and were intragastrically injected with 30 mg/kg tofacitinib (Fig. 5a), similar to the schedule used for the LPS-induced mice (Fig. 2a). Predictably, social interaction was significantly increased in tofacitinib-treated mice compared to untreated mice (Fig. 5b). These results revealed that tofacitinib treatment effectively improved social avoidance in CSDS-induced mice in the absence of any weight changes (Fig. 5c). Compared with untreated mice, tofacitinib-treated mice also showed a reduction in immobility time in the FST and TST, while locomotor activity in the OFT was significantly increased (Fig. 5d–f). We also employed the splash test to further evaluate the anhedonia of CSDS-induced mice. We found that tofacitinib treatment could significantly increase grooming time of CSDS-induced mice, indicating the anti-anhedonic effect of tofacitinib (Supplementary Fig. 5a, b). In the hippocampus, we observed a significant increase in the phosphorylation of JAK3, STAT1, and STAT3, while no notable changes in the phosphorylation of JAK1 or STAT5 were detected in untreated CSDS-induced mice (Fig. 5g–l). As expected, tofacitinib treatment markedly decreased the phosphorylation of JAK3, STAT1, and STAT3 in the hippocampus of CSDS-induced mice (Fig. 5i–k). Moreover, this pharmacological treatment effectively attenuated neuroinflammatory cytokine transcription and the activation of microglia in the hippocampus of CSDS-treated mice (Fig. 5m–o and Supplementary Fig. 6a–c). Moreover, these mice treated with tofacitinib were less susceptible to the CSDS-induced reduction in BDNF levels in the hippocampus (Fig. 5q, r), similar to the findings in LPS-induced mice treated with tofacitinib (Fig. 3j–l). These results suggested that tofacitinib could effectively prevent depressive-like behaviors in both LPS-induced and CSDS-induced depressive mice, partially via decreased hippocampal microgliosis and increased BDNF levels.

Fig. 5. The JAK inhibitor tofacitinib represents the similar effect on depression-related behavioral improvement in CSDS-induced mice.

a Schematic timeline of CSDS modeling, tofacitinib treatment, and behavioral tests. b Social index and time in interaction zone in the social interaction test (n = 9–16). c Relative body weights differences (n = 9–16). d Immobility time in the forced swimming test, the immobility time in the last 4 min was analyzed (n = 9–16). e Immobility time in the tail suspension test, the immobility time in the last 4 min was analyzed (n = 9–16). f Total distance traveled in the open field test (n = 9–16). g representative blots showing JAKs and STATs expression (n = 3). h–l Representative bar graph of JAKs and STATs expression analysis (n = 3). m Representative immunofluorescence of Iba-1 in the hippocampus. n Column graphs representing Iba-1 expression (n = 3). o Analysis of CCL2, IL-1β, and TNF-α mRNA levels in the hippocampus (n = 6). p Analysis of BDNF mRNA levels in the hippocampus (n = 6). q Representative blots showing BDNF expression (n = 3). r Representative bar graph of BDNF expression analysis (n = 3). Data are expressed as mean ± SEM, *P < 0.05, **P < 0.01.

Tofacitinib inhibits pro-inflammatory cytokines in microglia and increases BDNF levels in neurons in vitro

To test the direct effect of tofacitinib on microglia, BV2 cells (a microglial cell line) were introduced in vitro study (Fig. 6a). We first assessed the cytotoxicity of tofacitinib on BV2 cells by the CCK8 assay, revealing that tofacitinib had no toxic effect at concentrations up to 75 μM (Fig. 6b). Thus, we chose the 75 μM concentration for further investigations. In vivo depressive status or in vitro LPS stimulations, microglia could significantly increase pro-inflammatory cytokines and subsequently impair neuronal functions, which might be reversed by tofacitinib. To test this hypothesis, BV2 cells were treated with 75 μM tofacitinib and stimulated by 100 ng/ml LPS for 24 h. BV2 cells were next lysed to analyze mRNA expression, and their supernatants were removed to HT22 cell (a hippocampal cell line) to evaluate changes in BDNF levels (Fig. 6c). Consistent with findings in vivo, we found that tofacitinib treatment in LPS-induced BV2 cells significantly decreased the mRNA of pro-inflammatory cytokines, such as CCL2, IL-1β, and TNF-α (Fig. 6d). As expected, the supernatants with Tofa-treatment could effectively increase BDNF levels in HT22 cells, compared to those with non-treatment (Fig. 6e, f). These results demonstrated that tofacitinib could indeed inhibit hippocampal microgliosis and increase BDNF levels in vitro and vivo, contributing to its anti-depressant effects. Overall, our study indicated that the JAK inhibitor tofacitinib has potential translational implications for depression prevention beyond its current indications.

Fig. 6. The JAK inhibitor tofacitinib suppresses microglial cytokines and increases neuronal BDNF levels in vitro.

a Schematic diagram of the CCK8 assay on BV2 cells. b The assessment of the toxic effect of tofacitinib on BV2 cells (n = 5). c Schematic diagram of the protective effect of tofacitinib on BV2 cells and HT22 cells. d Analysis of CCL2, IL-1β, and TNF-α mRNA levels in BV2 cells (n = 4). e Representative blots showing BDNF expression (n = 3). f Representative bar graph of BDNF expression analysis (n = 3). Data are expressed as mean ± SEM, *P < 0.05, **P < 0.01.

Discussion

Previous studies have shown that inflammatory levels are relevant to depressive disorder, especially at the onset of depression. The JAK-STAT pathway is critical for inflammatory signaling and is involved in several neurological disorders [36, 37]; however, its role in depression is poorly understood. Thus, we hypothesized that the JAK-STAT pathway could be implicated in the depressive disorder observed in neuroinflammatory contexts and beyond. In the present study, we revealed that aberrant activation of the JAK-STAT pathway in the hippocampus is highly conserved in LPS-induced and stress-induced depressive mouse models. In light of these findings, we utilized the JAK inhibitor tofacitinib, an FDA-approved drug, to prevent depressive-like behaviors in both depressive models. In addition, the antidepressant effect of tofacitinib was attributed to the prevention of neuroinflammation and increased BDNF levels in the hippocampus.

A new concept proposed that inflammation is a potential cause of depression [38]. Research has indicated that inflammation affects approximately 1/4 of patients with depression and is linked to a poor prognosis and treatment resistance [39]. The LPS-induced depression model is a widely used approach for studying the connection between inflammation and depression [40, 41]. Lipopolysaccharide (LPS), found in the cell membrane of gram-negative bacteria, triggers the production of proinflammatory cytokines in the brain and periphery [42]. LPS could activate TLR4, via TRIF and MyD88-dependent ways to trigger the classical MAPK family and the NF-κB pathway. Subsequent researches have revealed that inflammatory responses stem from a sophisticated intracellular signaling network, where the interaction between JAK-STATs and p38MAPK holds significant importance [43]. Additionally, JAK/STAT pathway activation also was reported as a positive NF-κB feedback loop [44]. Eventually, this complex intracellular signaling network could increase the production of numerous pro-inflammatory mediators, such as TNF-α, IL-1β, and CCL2 [45, 46], which concurs with our results in vitro. In rodent studies, LPS exposure leads to a sequence of sickness behaviors, such as reduced body weight and food intake, followed by depression-like behaviors [47]. Systemic LPS administration in mice increases the brain levels of proinflammatory cytokines, resulting in lethargy, reduced locomotor activity, and sleep disturbances [47, 48]. Interestingly, anxiety-like behavior also emerges during this phase. Approximately one day after LPS injection, mice exhibit depressive-like behavior, characterized by increased immobility duration in the FST and TST [48, 49]. Our results concur with earlier research indicating that LPS induces depressive-like behavior in the FST and TST but does not affect locomotor activity in the OFT.

In addition to inflammation-driven depression, we also employed the chronic social defeat stress-induced depressive model, which exhibits a range of negative behaviors similar to those of humans [50]. In everyday life, individuals often face various stimuli that arise from interactions with others. These interactions can sometimes be challenging and stressful, not only for humans but also for social animals [51, 52]. When mice are subjected to repeated social defeat stress, depressive-like behavior occurs [53]. In our research, we observed similar patterns in mice subjected to lipopolysaccharide or chronic social defeat stress. These mice demonstrated an increase in immobility during the TST and FST. These behaviors suggest a state of behavioral despair.

Clinical trials have shown promising results regarding the antidepressant effects of anti-inflammatory treatment on depression [54–56]. However, it is debated whether a specifical inflammatory pathway plays a critical role in depression, and mechanism-based drug discovery is urgently needed. Although anti-inflammatory treatment shows promise as a potential avenue for managing depression in patients [57], the underlying mechanism is still being investigated, providing an opportunity to explore the inflammatory effects of the JAK-STAT pathway [58]. The JAK-STAT pathway is implicated in various physiological processes, including inflammation and stress-related diseases, which are linked to the onset of depression [59]. In this study, we comprehensively assessed the phosphorylation of JAKs and STATs, which are indicators of JAK-STAT activation, in LPS-induced and stress-induced depressive mouse models. Importantly, aberrant activation of the JAK-STAT pathway is common and conserved in both depressive mouse models. In line with our findings, anti-inflammatory treatments, such as erythropoietin (EPO), have been shown to prevent neuroinflammation and relieve depression via the JAK/STAT pathway partially at least [48]. A recent study also suggested that the activation of JAK-STAT signaling is an important pathway underpinning immune disorders in major depression by the protein–protein interaction network analysis [15]. However, it is challenging to isolate contributions of a specific JAK to depressive disorders. Our results suggest that the antidepressant effect of tofacitinib is closely related to inhibition of JAK3 activity in the hippocampus. Consistent with our findings, clinical data analysis has also shown that the transcription and protein levels of JAK3 are significantly increased in patients with depression [60]. However, systematic inhibition of JAK3 could dysregulate immune system. To minimize this side effect, future work should develop nervous-targeted JAK3 inhibitors for treating depression.

Although clinical studies have shown the association between JAK/STAT pathway and depression, their causality still is not validated by pharmacological manipulation, especially in classic depressive models. To investigate whether pharmacological suppression of the JAK-STAT signaling pathway has an effect on improving depression, we chose the marketed drug tofacitinib to examine its ability to ameliorate depressive-like behavior in animal models of depression. The pan-JAK inhibitor tofacitinib, developed by Pfizer, is a JAK kinase inhibitor that effectively inhibits the activity of JAK1 and JAK3, blocking the signal transduction of various inflammatory factors [61]. Recently, clinical studies have confirmed that tofacitinib has the ability to cross the blood‒brain barrier and has the potential to treat mood disorders [20]. In normal states, we found that tofacitinib does not affect behaviors in normal mice but improves behaviors in depressive animals. This interesting phenomenon could be explained through several biological mechanisms. On the one hand, normal mice have lower levels of JAK activation compared to depressive animals in the hippocampus, where this pathway might be upregulated due to chronic inflammation. On the another, depression is associated with increased levels of pro-inflammatory cytokines and activation of the immune system in the brain [62, 63]. This inflammatory response can contribute to the behavioral symptoms of depression. In contrast, normal mice typically have a stable, low level of inflammation under standard conditions. Tofacitinib’s impact on behavior could thus be related to its ability to attenuate this heightened inflammatory response in depressive animals, leading to behavioral improvements.

In the depressive states, we found that tofacitinib can prevent an increase in immobility time, an indicator of antidepressant-like activity in both the FST and TST, two established behavioral paradigms of depression in mice. Interestingly, 30 mg/kg tofacitinib had a greater effect on the prevention of depressive-like behaviors than did 70 mg/kg tofacitinib, which warrants further investigation. The antidepressant-like effects of tofacitinib may be related to the prevention of neuroinflammation and the increase in BDNF levels. In both depressive models, tofacitinib treatment ameliorated STAT-mediated cytokine transcription and reduced the number of microglia in the hippocampus. This finding is in accordance with the notion that microgliosis in the hippocampus is strongly associated with the progression of depression in humans [9, 64]. In addition, decreased BDNF levels in the hippocampus have also been observed in mouse models and patients with depression [65, 66]. Our results demonstrated that tofacitinib treatment prevented the reduction in BDNF levels in the hippocampus of mice exposed to LPS or CSDS. Similarly, other anti-inflammatory agents have been shown to increase BDNF in preclinical models of depression [67, 68]. These data indicate that the prevention of hippocampal microgliosis and decreased BDNF contribute to the antidepressant-like effect of tofacitinib. Further clinical studies are necessary to investigate the potential of tofacitinib as a new option for the management of depression.

Materials and methods

Experimental animals

All C57BL/J6 mice were housed under standard laboratory conditions (22 ± 1 °C, 55% ± 5% humidity) with a 12:12 h light/dark cycle and food and water provided ad libitum. Twelve-month-old male CD-1 mice were used for the CSDS model and fed alone. Eight-week-old wild-type male mice were used for the CSDS model and subsequent tofacitinib treatment experiments. Twelve-week-old wild-type male mice were used for the LPS model and subsequent tofacitinib treatment experiments. All animal experiments and protocols were approved by the Animal Care and Use Committee of the Shanghai Institute of Materia Medica, where the experiments were conducted.

Experimental design for drug treatment

In the first experiment, the experimental animals were divided into two groups (each group = 7–10 animals): the vehicle saline-treated group (Veh) and the lipopolysaccharide-induced depressive-like group (1 mg·kg⁻¹·d⁻¹) (LPS, MedChemExpress, HY-D1056). LPS was administered intraperitoneally (0.1 ml/10 g mice). The drug treatment schedule is shown in Fig. 1a. In the second experiment, the experimental animals were divided into three groups (each group = 7–10 animals): the Veh group, the LPS group, and 30 mg/kg tofacitinib (Tofa, MedChemExpress, HY-40354) plus LPS group (TO + LPS). Tofacitinib was diluted in 0.5% carboxy methyl cellulose solution, and tofacitinib was intragastrically injected (0.1 mL/10 g mice) 1 h before daily LPS administration. The drug treatment schedule is shown in Fig. 2a. In the third experiment, the experimental animals were divided into two groups (each group = 10–16 animals): the Veh group and the CSDS group. The experimental schedule is shown in Fig. 3a. In the fourth experiment, the experimental animals were divided into three groups (each group = 10–16 animals): the Veh group, the CSDS group, and the tofacitinib (30 mg·kg⁻¹·d⁻¹) plus CSDS group (TO + CSDS). The drug treatment schedule is shown in Fig. 4a.

Chronic social defeat stress (CSDS) procedures

The procedures for CSDS were established according to published reports [22, 50]. Male CD-1 mice were screened for 3 consecutive days to ensure their aggressiveness. For 10 consecutive days, except for the nonstressed vehicle-treated mice, the C57BL/6J mice were individually placed as “invaders” in aggressive CD1 mouse cages daily for 10 min. After physical defeat, a perforated divider separated them for 24 h until the next physical defeat. Each day, the intruder mice always faced a different resident. Likewise, two mice from the control group were kept in the breeding cage separated by a perforated divider.

Behavioral assays

All behavioral tests were performed during the light cycle after the participants had handled for at least 3 days. Mice were habituated to the behavior room for at least 1 h before testing. Mouse movements were recorded using a video tracking system (SuperMaze video tracking software, XinRuan Information Technology, Shanghai, China). The behavior box was cleaned with 70% ethanol after each trial to eliminate any olfactory cues.

Social interaction test (SIT)

In the first stage, the experimental mice were allowed to explore an open field arena (40 cm × 40 cm × 40 cm) with an empty plastic box (10 cm × 7 cm × 18 cm) for 2.5 min. In the second stage, an unfamiliar CD-1 mouse was placed in the plastic box, the experimental mouse was placed in the arena again, and activity was recorded for another 2.5 min. The time that the mice spent in the interaction zone surrounding the plastic box was recorded and analyzed. For the SI test, a social index was calculated. Social index = time spent in the interaction zone with a CD1 mouse/time spent in the interaction zone without a CD1 mouse. Susceptible mice were defined by an SI < 1, whereas resilient mice were defined by an SI > 1.

Forced swim test (FST)

Mice were placed individually into a transparent plastic cylinder (diameter 15 cm, height 30 cm), filled with 15 cm depth of water (23 ± 1 °C). The test time was 6 min, and the immobility time of each mouse was scored over the last 4 min by an investigator blinded to the groups. Immobility was defined as floating or a lack of active movement. For each trial, the water was replaced.

Tail suspension test (TST)

Mice were suspended by the tails by using adhesive tape 1 cm from the tip of the tail and 15 cm above the ground. The tests lasted for 6 min, and the immobility time was measured and analyzed during the final 4 min. Immobility was defined as the absence of movements of the animal’s head and body.

Open field test (OFT)

We used a square box (40 cm× 40 cm× 40 cm) and placed the mice in the center of the box. Mice were allowed to freely move for 5 min, the total distance traveled was recorded and analyzed using XinRuan software.

Splash test

The splash test was conducted on both control and CSDS-induced mice. Control animals were housed individually for one day before testing, while stressed mice had their cages changed simultaneously. This test was conducted under red light conditions (15 W). A spray atomizer containing a 10% sucrose solution was applied to the dorsal coat of each mouse in its home cage. The sucrose solution stained the coat and triggered grooming behavior. Following application of the sucrose solution, the duration of grooming behavior was recorded over a 5-min period.

Western blotting

Hippocampal tissues were quickly dissected from the mice after sacrifice. Tissue samples were homogenized in cold RIPA lysis buffer supplemented with phosphatase and protease inhibitors and lysed for 20 min. The protein concentration was determined using a standard BSA method (SB-WB013, Share-bio, Shang Hai). The proteins in the samples were separated on 10% SDS polyacrylamide gels (SKV-005, Share-bio, Shang Hai) and then transferred to polyvinylidene fluoride membranes. The membrane was blocked with 5% nonfat milk. The proteins were incubated with rabbit anti-JAK3 (1:1000, Thermo Fisher Scientific), rabbit anti-p-jak3 (1:1000, Cell Signaling Technology), rabbit anti-JAK1 (1:1000, Cell Signaling Technology), rabbit anti-p-JAK1 (1:1000, Abcam), rabbit anti-STAT1 (1:1000, Cell Signaling Technology), rabbit anti-p-STAT1 (1:1000, Cell Signaling Technology), rabbit anti-STAT3 (1:1000, Cell Signaling Technology), rabbit anti-STAT5 (1:1000, Cell Signaling Technology), rabbit anti-p-STAT5 (1:1000, Cell Signaling Technology), rabbit anti-p-STAT5 (1:1000, Cell Signaling Technology) and rabbit anti-α-tubulin (1:5000, ABclonal) antibodies overnight at 4 °C. After incubation with a specific primary antibody overnight, a horseradish peroxidase-conjugated secondary antibody was applied. In some cases, membrane stripping buffer (SB-WB007, Share-bio, Shang Hai) was employed to incubate another antibody. The signals were visualized using enhanced chemiluminescence (ECL) reagent (SB-WB011, Share-bio, Shanghai, China).

RNA isolation and gene expression analysis

RNA was isolated from the hippocampus using the TRIzol and chloroform extraction methods. The concentration and purity of the RNA were detected with a NanoDrop 2000. The Prime Script RT Master Mix (ABclonal, RK20402) was used to reverse transcribe 1 μg of total RNA into cDNA. According to the manufacturer’s instructions, real-time qPCR was carried out on a QuantStudio5 instrument (Applied Biosystems) using the SYBR Green QPCR Kit (ABclonal, RK21203). Quantitative real-time polymerase chain reaction (PCR) was performed, and the threshold amplification cycle number (Ct) was determined for each reaction in the linear phase of the amplification plot. The normalization control for relative quantification of the data was either GAPDH or β-actin. The sequences of primers used were as follows: β-actin, forward: 5’-AGTGTGACGTTGACATCCGT-3’, reverse: 5’-GCAGCTCAGTAACAGTCCGC-3’; GAPDH, forward: 5’- AGGTCGGTGTGAACGGATTTG-3’, reverse: 5’- TGTAGACCATGTAGTTGAGGTCA-3’; TNF-α, forward: 5’- CCCTCACACTCAGATCATCTTCT-3’, reverse: 5’- GCTACGACGTGGGCTACAG-3’; CCL-2, forward: 5’-GCTACAAGAGGATCACCAGCAG-3’, reverse: 5’-GTCTGGACCCATTCCTTCTTGG-3’; IL-1β, forward: 5’-GCAACTGTTCCTGAACTCAACT-3’, reverse: 5’-ATCTTTTGGGGTCCGTCAACT-3’; and BDNF, forward: 5’- GGCTGACACTTTTGAGCACGTC-3’, reverse: 5’- CTCCAAAGGCACTTGACTGCTG-3’. The data are reported as the fold increase in mRNA levels in the treated samples relative to the control.

Immunofluorescence

After the behavioral tests, the test mice were anesthetized and perfused transcardially with 0.1 M phosphate buffer. The brains were separated, postfixed, and dehydrated in 30% sucrose solution. Hippocampal sections (30 µm) were collected using a freezing microtome. Then, the slices were washed in PBS for 5 min three times and blocked with 5% bovine serum albumin (BSA) and 0.2% Triton X-100 for 1 h at room temperature. After blocking, the sections were incubated with a rabbit primary antibody against IBA1 (1:250; Servicebio) overnight at 4 °C. After washing, the sections were sequentially incubated with secondary antibodies for 1 h at room temperature. The nuclei were counterstained with DAPI for 10 min. Fluorescence images were captured with a fluorescence microscope. The number of cells in the intact hippocampus of 3 mice in each group was measured using ImageJ.

Cell culture and treatment

Mouse hippocampal cell line HT22 cells and mouse microglial cell line BV2 (ATCC, Rockville, MD, USA) were cultured in DMEM medium (Thermo Fisher Scientific) containing 10% fetal bovine serum (Gibco) and 1% penicillin/streptomycin under a humidified atmosphere of 5% CO₂ in air at 37 °C. The medium was refreshed daily, and cells were passaged upon reaching 80% confluence. To assess the effect of tofacitinib on BV2 and HT22 cell, BV2 cells were cultured on 24 well and pretreated with 75 μM tofacitinib for 1 h, followed by stimulation with 100 ng/ml LPS for 24 h. BV2 cells next were lysed to mRNA analysis (SB-R001, Share-bio, Shang Hai) and their medium supernatants were moved to HT22 cells. HT22 cells were cultured with BV2’s medium supernatants for 6 h and then were lysed by SDS-loading (SB-PR037, Share-bio, Shang Hai) to assess BDNF levels.

The assessment of cell viability

To evaluate the cytotoxic effects of tofacitinib and determine cell viability, a CCK-8 assay was conducted following the manufacturer’s instructions. Initially, cells were seeded into a 96-well plate and allowed to culture for 12 h. Subsequently, the culture medium was replaced with fresh medium containing varying concentrations of tofacitinib. After 24 h of incubation, 10 μL of CCK-8 solution (SB-CCK8, Share-bio, Shang Hai) was added to each well and incubated for an additional 1 h. The absorbance at 450 nm was measured using a microplate spectrophotometer reader.

Quantification and statistical analysis

All the data are representative of two or three independent experiments. All the data are presented as treatment means ± SEM and were analyzed with commercially available GraphPad Prism software (GraphPad, Inc.). The data were analyzed with two-tailed unpaired Student’s t tests or one-way ANOVA followed by Fisher’s least significant difference (LSD) tests with two-tailed distributions. Statistical significance was determined at *P < 0.05, **P < 0.01.

Author contributions

YNG and KJP conceived the study, designed and conducted experiments, analyzed results, and contributed to the manuscript preparation. YMZ and YBQ contributed to writing review and editing. HRG, TY, WGC, TZ, HCZ, JWZ, XCL, ZTC, and ZC contributed to the performance of experimental and behavioral procedures. YZ and JL supervised the project. All authors have read and approved the article.

Data availability

This paper does not report any original code. All data generated and supporting the findings of this study are available.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s41401-024-01384-8.