Neuroimmune Responses Mediate Depression-Related Behaviors following Acute Colitis.

Many patients with visceral inflammation develop pain and psychiatric comorbidities such as major depressive disorder, worsening the quality of life and increasing the risk of suicide. Here we show that neuroimmune activation in mice with dextran sodium sulfate-induced colitis is accompanied by the development of pain and depressive behaviors. Importantly, treatment with the flavonoid luteolin prevented both neuroimmune responses and behavioral abnormalities, suggesting a new potential therapeutic approach for patients with inflammatory bowel diseases.

Introduction

Gastrointestinal inflammation during inflammatory bowel diseases (IBDs) involves the activation of immune responses that lead to the release of inflammatory mediators, such as cytokines and chemokines (Strober and Fuss, 2011). This, in turn, gives rise to recurrent visceral pain and causes the development of psychiatric comorbidities such as mood disorders (Walker et al., 2013). Indeed, up to 80% of patients with IBD suffer from anxiety symptoms and 60% develop major depressive disorder (MDD) that often outlasts the resolution of pain symptoms and leads to an increased risk of suicide (Chen et al., 2015). These comorbidities have a major detrimental impact on the quality of life of these patients, but the underlying cellular and molecular mechanisms have not been determined. Although symptoms of depression can potentially be treated with antidepressants (Murrough et al., 2017), a clear unmet need is the ability to prevent the development of psychiatric disorders in patients suffering from inflammatory conditions affecting the gastrointestinal tract.

Functional magnetic resonance imaging studies in patients with IBD have revealed changes in the activity of a number of different brain regions, including the prefrontal cortex (PFC) and hippocampus (Ma et al., 2015, Bao et al., 2016). Among these, the hippocampus is a key brain region within the limbic system that has been linked to depression in both humans and rodents (Dantzer et al., 2008), and during depressive states, there is a dysregulation of several neurotransmitter systems such as the serotonergic, noradrenergic, and GABAergic systems. mRNA levels of specific inflammatory markers, such as for COX-2, are upregulated in the hippocampus of mice with visceral inflammation (Do and Woo, 2018), and prolonged pharmacologically induced visceral inflammation gives rise to reduced hippocampal neurogenesis (Zonis et al., 2015). In addition, there is an inflammatory component to depression, as evident from alterations of interleukin 1 beta (IL-1β) and tumor necrosis factor alpha (TNF-α) levels in the limbic system of patients with depression (Hodes et al., 2015). Furthermore, anti-inflammatory drugs can be effective in treating comorbid depression-like symptoms in patients with inflammatory disorders (for review see, Hodes et al., 2015). In rodents, both IL-1β (Ikegaya et al., 2003) and TNF-α (Batti and O'Connor, 2010) have been shown to alter synaptic plasticity in CA1 pyramidal cells.

We therefore reasoned that during visceral inflammation, central nervous system action of inflammatory cytokines released by peripherally activated leukocytes may give rise to alterations in brain function, which in turn lead to neuropsychiatric consequences such as depression. To examine this issue, we employed the dextran sodium sulfate (DSS) model of visceral inflammation. This model is pertinent as it primarily involves the innate immune response and is characterized by inflammatory cell infiltration (along with weight loss, bloody diarrhea, epithelial cell damage) and increased production of inflammatory mediators such as TNFα, IL-6, IL-1, and interferons (primarily the innate immune response) (Lopes et al., 2018). We find that mice treated with DSS develop visceral inflammation and depression-like behaviors. Mice undergoing DSS treatment display immune cell infiltration into the brain microvasculature that is accompanied by an elevation of IL-1β levels. The natural flavonoid luteolin prevents both the neuroimmunological events and the development of depression.

Results

We used a 6-day treatment of DSS in drinking water as an experimental model of inflammatory colitis in mice. This model was chosen because DSS disrupts the epithelial barrier, which causes mucosal infection from commensal bacteria and thus colonic inflammation (Lopes et al., 2018). Compared with vehicle-treated animals, DSS mice exhibited reduced body weight over the study time course of 1 month (Figure S1A) and significant neutrophil migration to the colon up to 2 weeks following DSS discontinuation, as revealed by a myeloperoxidase assay (Figure S1B). Mice treated with DSS developed visceral hypersensitivity (see below) that was evident for a week following DSS treatment. Different cohorts of mice were treated with DSS and then tested for depression-related behaviors. We focused predominantly on two time periods. One set of experiments was designed to examine the early responses to DSS during the inflammatory phase, including measurements of gut permeability; measurements of leukocyte infiltration, which is expected to occur during the inflammatory phase; visceral hypersensitivity, which is known to resolve before the 4-week time point (Costa et al., 2012); and weight gain as a function of time after DSS treatment (up to four weeks, see above). To examine the long-term effects of DSS on behavior and associated neurochemical changes, experiments on comorbidities with depression were typically conducted 4 weeks after the discontinuation of DSS, including all critical behavioral measurements for depression, the electrophysiology, and PCR measurements. Figure S2 depicts a summary of the time lines used in this study.

Onset of Depression-Related Behaviors in Male and Female Mice

In the novelty suppressed feeding test, a conflict paradigm test based on hyponeophagia, which is among the most commonly used test for screening of novel antidepressants, both male and female mice subjected to acute visceral inflammation displayed increased latency to interact with food (Figure 1A). Two-way ANOVA revealed a significant difference following DSS treatment (F(1,99) = 3.3; p = 0.0058) without sex differences in our observations. The effect of DSS treatment on the loss of self-care and motivational behavior was inferred by the time spent grooming in the splash test. The results show decreased time spent grooming for DSS-treated mice (Figure 1B). Two-way ANOVA revealed significant differences in response to DSS treatment (F(0,64) = 3.4; p < 0.0001) and for the sex versus DSS treatment interaction (F(1.88) = 11.10; p < 0.0179). Next, we verified whether DSS treatment affects the immobility time in models of depression-related behavior induced by despair. Both the forced swimming test (FST) and the tail suspension test (TST) are used to evaluate depression-related behaviors in rodents, with immobility time being decreased by classical antidepressants (Steru et al., 1985). DSS-treated mice exhibited a significantly elevated immobility time in both FST (Figure 1C) and TST (Figure 1D) compared with control groups for both sexes. Two-way ANOVA showed statistical differences in response to DSS treatment (F(1,06) = 3.4; p < 0.0001) and for the sex versus DSS treatment interaction (F(2,16) = 10.9; p < 0.0001) for FST. For the TST, a difference in response to DSS treatment was revealed (F(1,82) = 3.3; p = 0.0022). All groups of mice displayed normal ambulatory behavior in the open field when assessed at the same time points (between 3 and 4 weeks) as for the FST and TST with no differences found in the number of crossings for either DSS (H2O versus DSS) or sex (data not shown), thus suggesting that the depression-related behavior of DSS-treated mice was not due to motor deficits. The tricyclic antidepressant imipramine reversed the DSS effects in the TST (Figure 1E). Two-way ANOVA showed statistical differences following DSS treatment (F(3,82) = 8.8; p = 0.0125) and for the DSS versus imipramine treatment interaction (F(6,72) = 8.8; p < 0.0001). Collectively, our data show that mice with experimental colitis show a comorbid depression-like phenotype that is consistent with what is observed in patients with IBD. Because similar findings were obtained in both male and female mice, we focused on male mice for all ensuing experiments to reduce the numbers of mice required.

Figure 1

Development of Depression-Related Behaviors Following Acute Colitis

(A–D) (A) Novelty suppressed feeding test, (B) splash test, (C) forced swimming test, and (D) tail suspension test.

(E) Effect of imipramine (10 mg/kg, intraperitoneally) in male mice treated with DSS in the tail suspension test. Each bar represents the mean ± SEM, numbers depecited in the bars represent numbers of mice tested. The experimenter was blinded to the groups during splash test and forced swimming test.

Two-way ANOVA reveals behavioral abnormalities of mice treated with DSS; ap < 0.05, aaap < 0.001 DSS- versus H2O-treated groups; #p < 0.05, ###p < 0.001 male versus female mice interaction; ∗p < 0.05, ∗∗∗p < 0.001 imipramine- versus vehicle-treated mice.

Changes in Electrophysiological Properties and Gene Expression in the Hippocampus

We then performed an electrophysiological analysis of hippocampal neuron function using hippocampal slices. We measured the magnitude of evoked inhibitory post-synaptic currents (IPSCs) in CA1 pyramidal cells in response to electrical stimulation of Schaffer collaterals at a range of stimulus intensities. We found that IPSCs recorded from pyramidal neurons of DSS-treated animals 4 weeks after DSS discontinuation were significantly increased in magnitude compared with vehicle-treated animals (Figure 2). Two-way repeated measures (RM) ANOVA revealed statistical differences after DSS treatment (F(5,85) = 5.5; p = 0.0206). These data suggest the possibility that CA1 pyramidal cells may be under augmented inhibitory control and are consistent with the known involvement of the hippocampus in depressive states (Boddum et al., 2016).

Figure 2

Electrophysiological Properties of the Hippocampus Following Colitis

(A) Whole-cell patch clamp 4 weeks after end of DSS reveals that evoked Inhibitory Post Synaptic Current (eIPSCs) in pyramidal cells of the CA1 area were enhanced and (B) the eIPSC threshold is lower in mice following acute colitis. Each circle (left) and bar (right) represents the mean ± SEM. Two-way RM ANOVA shows that the feedforward inhibitory synaptic transmission is enhanced in Schaffer collateral input of hippocampal CA1 of mice following inflammatory colitis. ap < 0.05.

We then performed quantitative PCR analysis of hippocampal tissue at the same time point (4 weeks), and this experiment revealed an elevation of IL-1β mRNA (Figure 3A), but curiously no changes in TNF-α mRNA levels (Figure 3B). Nonetheless, the persistent upregulation of IL-1β levels suggest a sustained inflammatory process in the brains of DSS-treated mice and is consistent with previous reports implicating this cytokine in depression in rodents (Alcocer-Gómez et al., 2017) and humans (Kolaczkowska and Kubes, 2013). Possible microglia activation was characterized by flow cytometry. Activated microglia are identified as CD11b+CD45intermediate-expressing cells as described previously (Andonegui et al., 2018). Analysis of total brain samples showed no activation of microglial cells 1 week following DSS discontinuation (Figure 3C), thus indicating that the elevated IL-1β mRNA levels are not dependent on microglial activation.

Figure 3

Expression of Genes Related to Inflammation and Depression

(A and B) qPCR quantification of (A) IL-1β and (B) TNF-α mRNA expression from total hippocampus of mice isolated 4 weeks after discontinuation of DSS or vehicle treatment. Bars represent the mean ± SEM and are representative of two independent experiments. A t test reveals increased IL-1β mRNA expression in mice treated with DSS. *p < 0.05.

(C) Microglia CD11b+CD45 intermediate expression by flow cytometry 1 week after DSS discontinuation. Total brain cells from H2O- and DSS-treated mice were isolated and labeled. The CD11b+ cells were gated and analyzed for CD45 intermediate expression. The representative histogram shows that there is no increase in CD45 intermediate (CD45int) expression in the brains of DSS-treated mice (4.685 ± 0.459) when compared with H2O-treated animals (5.395 ± 0.060). Numerical values represent the mean ± SEM and are representative of three mice per group.

Altogether these data indicate that there are persistent changes in gene expression and functional properties in the hippocampus.

Live Imaging of Infiltrating Leukocytes into the Brain Microcirculation

Enhanced IL-1β in the brain of DSS mice and the lack of microglial activation inspired us to verify leukocyte recruitment in the brain microcirculation by using CX3CR1GFP/WT CCR2RFP/WT mice to track inflammatory and patrolling monocytes. These mice were additionally injected with antibody to label neutrophils. At 7 days after DSS discontinuation, there were no leukocytes in the brain microvessels of the PFC of control mice. However, DSS mice displayed a significant increase in neutrophils and both classical (CCR2hiCX3CR1low) and patrolling (CCR2lowCX3CR1hi) monocytes (Figure 4A, 4B, 4D, and 4E and Videos S1 and S2). Importantly, as there appeared to be no activation of microglia, this experiment further confirmed our flow cytometry analysis. Neutrophils mainly mediate tissue damage through activation by cytokines and release of oxidants, proteases, and other factors (de Oliveira et al., 2016). They are thought to re-enter the circulation through reverse neutrophil migration, thus potentially disseminating inflammation and causing damage far away from the original site of injury (Kolaczkowska and Kubes, 2013). The observation that an elevated number of neutrophils and monocytes were found in the brain microcirculation after inflammation of the gut supports the concept of long-lasting neuroimmune regulation of brain properties that may be correlated with the behavioral abnormalities found in DSS mice. Moreover, the inflamed blood-brain barrier (BBB) endothelium is known to exhibit increased cytokine expression such as IL-1β and TNF-α, which contribute to modulate the permeability of the BBB and the phenotype of infiltrating leukocytes (Shechter et al., 2013).

Figure 4

Imaging of Infiltrating Leukocytes into the Brain Microcirculation

(A–C) Intravital microscopy images of prefrontal cortex (PFC) vasculature performed 1 week after DSS discontinuation of (A) control (H2O), (B) colitis (DSS + vehicle), and (C) colitis group treated with luteolin (15 mg/kg, intraperitoneally daily, 15 days starting 2 days before DSS treatment) in reporter mice CX3CR1GFP/WTCCR2RFP/WT. Neutrophils are labeled in blue, CCR2 inflammatory monocytes are labeled in red, and CX3CR1 patrolling monocytes are shown in green.

(D) Quantification of total (left) and differentiated (right) rolling leukocytes.

(E) Quantification of total (left) and differentiated (right) adherent leukocytes. Neutrophils are visualized in blue as detected by anti-Ly6G monoclonal antibody (mAb) (shown by arrows), CCR2+ monocytes are visualized red (shown by arrows), endothelium is visualized in red by anti-CD31 mAb, and microglia are visualized in green. Each bar represents the mean ± SEM (n = 4–5). Two-way ANOVA reveals an increased presence of leukocytes in the brain vasculature of DSS-treated mice and protective effect of luteolin treatment. ap < 0.05, aap < 0.01, aaap < 0.001 DSS- versus H2O-treated mice; #p < 0.05, ##p < 0.01, ###p < 0.001 luteolin- versus vehicle-treated mice.

Luteolin Prevents Leukocyte Infiltration, Visceral Pain, and Depression-Related Behavior

Flavonoids are commonly present in a wide range of plants used in traditional medicine to treat a number of disorders and often exhibit anti-inflammatory properties (Spagnuolo et al., 2018). The bioflavonoid luteolin is found in many plants such as peppermint, artichokes, peppers, and carrots and is known to exhibit anti-oxidant, anti-microbial, anti-allergic, neuroprotective, anti-diabetic, and cardioprotective properties (López-Lázaro, 2009). Most importantly, its potent anti-inflammatory actions have been reported both in vitro and in vivo and appear to be related to its inhibitory action of transcription factors such as nuclear factor (NF)-κB, inducible nitric oxide synthase (iNOS), and inhibition of the protein kinase B and mitogen-activated protein kinase pathways (Aziz et al., 2018). We thus evaluated its effect on DSS-induced physiological changes in mice. Treatment of mice with luteolin (15 mg/kg, intraperitoneally, once daily for 15 days starting 2 days before the beginning of DSS treatment) abolished invasion of rolling (Figures 4C and 4D) and adhesion monocyte (Figures 4C and 4E) as well as rolling neutrophils, thus showing a near-complete prevention of leukocyte infiltration into brain microvasculature 1 week after DSS discontinuation (Video S3). Two-way ANOVA showed a significant difference due to DSS treatment (F(378,0) = 5.5; p = 0.0017) and for the DSS versus luteolin treatment interaction (F(6,96) = 5.4; p = 0.0341)) for rolling monocytes, and for adhesion monocytes (DSS treatment [F(101,0) = 4.4; p < 0.0001], and for the DSS versus luteolin treatment interaction [F(4,92) = 4.3; p = 0.0012)]). Furthermore, luteolin also completely prevented colonic hypersensitivity of DSS male mice 1 week post-DSS (Figures 5A and 5B, DSS treatment [F(4,47) = 9.9; p < 0.0001] and DSS versus luteolin treatment [F(5,40) = 9.9; p < 0.0001]) and eliminated depression-like behavior in the FST (Figure 5C, DSS treatment [F(1,02) = 7.12; p = 0.0008] and DSS versus luteolin treatment [F(1,78) = 9.12; p = 0.0021]) and TST (Figure 5D DSS treatment [F(1,91) = 5.5; p = 0.0051] and DSS versus luteolin treatment [F(5,98) = 5.5; p = 0.0069]), 4 weeks post-DSS. There was no effect on the number of crossings in the open field test (Figure S3), suggesting that locomotor dysfunctions are not major confounding factors in the interpretation of depression behaviors. We do acknowledge, however, that the FST and TST solicit abdominal muscles, which may be hypercontractile in visceral pain. Although the open-field test is not designed to detect such changes, we do note that there is likely no visceral hypersensitivity at the 4-week time point.

Figure 5

Luteolin Prevents Leukocyte Infiltration, Visceral Hypersensitivity, and Depression-like Behavior

Luteolin (15 mg/kg, intraperitoneally [i.p.] daily, 15 days) abolishes colonic hypersensitivity evaluated 6–7 days after DSS discontinuation in male mice.

(A–D) (A) Abdominal response and (B) area under the curve. Abdominal muscle contraction was recorded in response to distention pressures of 15, 30, 45, and 60 mm Hg. Each circle represents the mean ± SEM. (n = 10–11). Two-way RM ANOVA shows that luteolin completely abolishes colonic sensitivity of DSS-treated mice. ap < 0.05 DSS- versus H2O-treated mice; ∗p < 0.05 luteolin- versus vehicle-treated mice. Antidepressant-like effect of luteolin (15 mg/kg, i.p. daily, 15 days) following inflammatory colitis in male mice in the (C) FST and (D) TST. Bars represent the mean ± SEM and are representative of three experimental runs. Two-way ANOVA reveals antidepressant-like effect of luteolin. ap < 0.05, aaap < 0.001 H2O- versus DSS-vehicle mice; ∗∗p < 0.01, ∗∗∗p < 0.001 luteolin- versus vehicle-treated mice.

Luteolin Protects the Leaking Gut

Treatment of mice with DSS resulted in a decrease in transepithelial electrical resistance (TER), compared with controls 1 week from DSS discontinuation. TER is a widely used technique to quantify and determine the integrity of endothelial and epithelial tissues (Srinivasan et al., 2015). Luteolin treatment partially rescued the DSS-induced decrease in gut resistance of colitis mice with effect evident at 50 and 60 min. DSS-treated group exhibited decreased TER in colons mounted in Ussing chambers (Figure 6). Two-way RM ANOVA showed statistical difference in response to DSS treatment (F(4,17) = 12.12; p < 0.0001) and for the DSS versus luteolin treatment interaction (F(1,74) = 12.12; p < 0.0001)). These data thus suggest that the increased gut permeability during colitis can be partially reversed with luteolin.

Figure 6

Luteolin Reverses DSS-Induced Changes in Gut Permeability

Luteolin attenuates intestinal epithelial barrier dysfunction induced by DSS in mice. The distal colon was mounted in Ussing chambers after luteolin treatment (15 mg/kg, intraperitoneally daily, 15 days), and TER was measured for 60 min 1 week after DSS discontinuation. Circles represent mean ± SEM (n = 9–11). Two-way RM ANOVA reveals a significant effect of luteolin treatment. aap < 0.01, aaap < 0.001, aaaa p < 0.0001 H2O- versus DSS- treated groups; *p < 0.05 luteolin- versus vehicle-treated groups.

Discussion

Inflammation and the immune system have emerged as important components of MDD (Hodes et al., 2015) and are thought to serve as an induction mechanism of depression-related behaviors (Walker et al., 2013). Our data reveal that behavioral abnormalities also arise following acute inflammation of the gut and are associated with changes in electrophysiological properties of the hippocampus and activation of neuroimmune responses. The persistent upregulation of IL-1β levels observed in our study is consistent with multiple reports implicating this cytokine in depression and suggests a sustained neuroinflammatory process following peripheral inflammation that can be mediated by circulating leukocytes.

Although microglia are well characterized as an important source of IL-1β during neuroinflammation (Liu and Quan, 2018), our flow cytometry analysis suggests a different source other than the resident microglia in the brain for the elevated levels of IL-1β in the hippocampus. Our in vivo intravital microscopic studies revealed an increase in the infiltration of neutrophils and monocytes of both subtypes into cerebral microvasculature, which are consistent with the higher levels of IL-1β that we observed. After peripheral trauma or infection, neutrophils increase in number, and this is followed by an increase in monocytes that are recruited to the site of inflammation (Ginhoux and Jung, 2014). Neutrophil recruitment is initiated by changes in the surface of the endothelium as a result of stimulation by inflammatory mediators released from tissue-resident sentinel leukocytes. Neutrophils then facilitate the recruitment of monocytes into inflamed tissues (Kolaczkowska and Kubes, 2013). Monocytes might further contribute to inflammation but may also partake in its resolution (Ginhoux and Jung, 2014). These cells are able to re-enter the vasculature where they may be involved in spreading the inflammatory processes to other organs (Kolaczkowska and Kubes, 2013, de Oliveira et al., 2016), including immune-privileged sites, such as the brain (Shechter et al., 2013). The correlation between the peripheral immune system and depression is further supported by early studies showing increased neutrophils and ly6Chi monocytes in the blood of patients diagnosed with MDD (Smith, 1991, Maes, 1995) and by a more recent postmortem study that revealed an elevated number of peripheral monocytes in the brains of patients with depression (Torres-Platas et al., 2014).

The connection between depression and visceral inflammation appears to be bidirectional. In a mouse depression model based on reserpine-induced monoamine depletion, depression aggravated intestinal inflammation by altering tonic vagal inhibition of inflammatory cytokines (Ghia et al., 2008). Along these lines, depression induced by either bulbectomy or reserpine reactivates inflammation following chronic colitis (Ghia et al., 2009). Stress was also found to induce immune reactions and alter gut microbiota, thus sensitizing DSS-treated mice and modulating mood behaviors, along with increased infiltration of B cells, neutrophils, and proinflammatory macrophages in colonic lamina propria (Gao et al., 2018). On the other hand, acute colitis in rats induced by DSS produced both pain and anxious depressive behaviors when assessed 1–2 weeks following DSS discontinuation (Chen et al., 2015), which is qualitatively consistent with our findings presented here.

It is estimated that 30%–50% of patients with depression are refractory to treatment with classical antidepressants, thus it becomes evident that distinct mechanisms are involved in this pathology (Krishnan and Nestler, 2008) and highlights the necessity of alternative strategies for therapeutic interventions. In this context, bioflavonoids have emerged as natural sources for new therapeutics (Katiyar et al., 2012). Luteolin is a flavone present in medicinal plants, vegetables, and fruits, such as peppermint, artichoke, celery, broccoli, onions, and peppers among others (Aziz et al., 2018). Besides its anti-oxidant, anti-microbial, anti-diabetic, anti-allergic, and neuroprotective actions, luteolin has potent anti-inflammatory properties in vivo and in vitro mainly due to due to inhibition of nitric oxide production and the expression of nuclear factor NF-κB, which controls the production of inflammatory cytokines such as IL-1β (Seelinger et al., 2008, Aziz et al., 2018). Luteolin has been reported to be neuroprotective and to have antidepressant-like action when delivered orally to mice (Ishisaka et al., 2011). It also produced analgesia in an animal model of neuropathic pain when delivery spinally, but not supraspinally (Hara et al., 2014). Luteolin treatment did not inhibit neutrophil migration to the colon of DSS mice when analyzed 1 week following DSS discontinuation (data not shown); however, the intestinal anti-inflammatory action of luteolin was previously described in DSS-treated mice. Following systemic delivery, luteolin inhibited macrophage infiltration into the colonic mucosa, produced anti-inflammatory activity in the intestine, and improved disease activity index (DAI), such as colon shortening and histological damage assessments (Nishitani et al., 2013). It also significantly reduced DAI in the colon of DSS-treated mice; reduced the expression of mRNA for several pro-inflammatory mediators such iNOS, TNF-α, and IL-6; and increased the activities of colonic antioxidant factors such as Nrf2, SOD, and CAT (Li et al., 2016). These data are consistent with our findings showing that repeat dosing of systemically delivered luteolin prevents leukocyte infiltration into the brain, abolishes visceral hypersensitivity, and prevents the development of abnormal behaviors associated with depression.

Altogether, our findings reveal that acute visceral inflammation results in increased infiltration of leukocytes into the brain microvasculature and gives rise to depression-like behavior that is consistent with sustained neuroinflammation. Protecting the inflamed gut with the flavonoid luteolin prevents gut leakage and reverses behavioral abnormalities such as pain and depression. Thus the present study sheds light into the mechanisms that underlie the development of these comorbid conditions and suggest new ways for therapeutic intervention for the treatment of depression associated with visceral inflammation.

Limitations of the Study

It is important to recognize that the DSS model does not entirely reflect the etiology of human colitis (Chassaing et al., 2014). A second caveat is that although we examine changes in hippocampal function at the cellular level, the intravital microscopic experiments do not allow access to deep brain structures such as the hippocampus, and instead are conducted in the PFC. We believe that the findings obtained in this brain structure are also relevant to hippocampal function.

Methods

All methods can be found in the accompanying Transparent Methods supplemental file.