Paradoxical Effect of LTB4 on the Regulation of Stress-Induced Corticosterone Production

Depression is a mental illness with a complex and multifactorial etiology, which has been associated with stress and inflammation. Infections, autoimmune diseases, envenomation, and trauma induce an inflammatory response that is characterized by increasing levels of circulating cytokines (e.g., IL-1β) and lipid mediators [e.g., PGE2 and leukotrienes B4 (LTB4)]. Recently, we showed that LTB4 production by the 5-lipoxygenase (5-LO) pathway regulates IL-1β and PGE2 release, reducing tissue damage in a model of sterile inflammation. Since IL-1β and PGE2 increase in serum of stressed patients and potentially trigger depression, we used an animal model of chronic unpredictable stress (CUS) to investigate the potential impact of LTB4 over depression-like symptoms. At basal conditions, 5-LO deficiency (Alox5−/−) reduces the preference for sucrose, while inducing a higher immobilization time on the tail suspension test when compared 129sv. Moreover, Alox5−/− mice present increased caspase-1 expression and elevated levels of IL-1β, IL-17 and PGE2 in the spleen, with increasing corticosterone levels in the frontal cortex but reducing systemic levels. Compared to 129sv mice, CUS induced higher levels of systemic, frontal cortex and hippocampal corticosterone, and also reduced sucrose preference, increased levels of splenic IL-1β, IL-17 and PGE2 and reduced levels of LTB4. Interestingly, CUS exposure did not alter the reduced sucrose preference shown by Alox5−/− mice but greatly enhanced splenic PGE2 production. Compared to Alox5−/− mice at basal conditions, CUS exposure also increased levels of systemic corticosterone, which remained lower than those of CUS-129sv animals. We also observed that treatment with LTB4 decreased caspase-1 expression and systemic levels of corticosterone in CUS-Alox5−/− mice but there was no significant impact on the reduced sucrose preference. Our results demonstrate that LTB4 controls the hypothalamic-pituitary-adrenal (HPA) axis by regulating levels of systemic corticosterone associated with the repression of caspase-1 expression and production of inflammatory mediators. One limitation of our study is that 129sv and Alox5−/− mice were not littermates, not sharing, therefore, the same intra-uterine and preweaning environment. Even so, taken together our results indicate that 5-LO activity is critical for the regulation of stress-induced symptoms, suggesting that the Alox5−/− mouse could be a natural model of corticosterone-independent reduced reward sensitivity.

Depression is a mental illness with a complex and multifactorial etiology, which has been associated with stress and inflammation. Infections, autoimmune diseases, envenomation, and trauma induce an inflammatory response that is characterized by increasing levels of circulating cytokines (e.g., IL-1β) and lipid mediators [e.g., PGE 2 and leukotrienes B 4 (LTB 4 )]. Recently, we showed that LTB 4 production by the 5-lipoxygenase (5-LO) pathway regulates IL-1β and PGE 2 release, reducing tissue damage in a model of sterile inflammation. Since IL-1β and PGE 2 increase in serum of stressed patients and potentially trigger depression, we used an animal model of chronic unpredictable stress (CUS) to investigate the potential impact of LTB 4 over depression-like symptoms. At basal conditions, 5-LO deficiency (Alox5 −/− ) reduces the preference for sucrose, while inducing a higher immobilization time on the tail suspension test when compared 129sv. Moreover, Alox5 −/− mice present increased caspase-1 expression and elevated levels of IL-1β, IL-17 and PGE 2 in the spleen, with increasing corticosterone levels in the frontal cortex but reducing systemic levels. Compared to 129sv mice, CUS induced higher levels of systemic, frontal cortex and hippocampal corticosterone, and also reduced sucrose preference, increased levels of splenic IL-1β, IL-17 and PGE 2 and reduced levels of LTB 4 . Interestingly, CUS exposure did not alter the reduced sucrose preference shown by Alox5 −/− mice but greatly enhanced splenic PGE 2 production. Compared to Alox5 −/− mice at basal conditions, CUS exposure also increased levels of systemic corticosterone, which remained lower than those of CUS-129sv animals. We also observed that treatment with LTB 4 decreased caspase-1 expression and systemic levels of corticosterone in CUS-Alox5 −/− mice but there was no significant impact on the reduced sucrose preference. Our results demonstrate that LTB 4 controls the hypothalamic-pituitary-adrenal (HPA) axis by regulating levels of systemic corticosterone associated with the repression of caspase-1 expression and production of inflammatory mediators. One limitation of our study is that 129sv and Alox5 −/− mice were not littermates, not sharing, therefore, the same intra-uterine and preweaning environment.

INTRODUCTION
Depression is a serious public health problem and one of the most common psychiatric disorders in Western countries. It may lead to suicide, a tragic fatality accounting for mortality of 800,000 people every year (World Health Organization (WHO), 2017). Depressed patients often exhibit symptoms such as sadness, loss of interest or pleasure (anhedonia), feelings of guilt or low self-esteem, disturbances in sleep or appetite, low energy, and a reduced ability to concentrate (Zunszain et al., 2011). An association between stress, immune response, and psychological disorders has been proposed over the last few decades (Guan et al., 2014). A considerable number of evidence demonstrates that multiple aspects of the immune response are altered during stress or depressive episodes (Maes et al., 1997;Tymen et al., 2013). Indeed, the immune response has been shown to contribute to the pathophysiology of these disorders (Irwin and Miller, 2007). In conditions of chronic psychological stress, such as major depressive disorders, both the cellular and humoral immunity can be persistently activated and contribute to depression.
Recently, Zhang et al. (2015) reported that mice subjected to chronic mild stress (CMS) have increased IL-1β levels, caspase-1 activity, and NLRP3 inflammasome activation in the hippocampus. Pharmacological inhibition of the inflammasome decreased both IL-1β levels and the CMS-induced depressive-like behavior. Stress induced by pathogenic or non-pathogenic (sterile) stressors triggers the NLRP3 inflammasome in myeloid cells, resulting in the activation of caspase-1, which cleaves the precursor forms of IL-18 and IL-1β cytokines, leading to release of the active cytokines (Strowig et al., 2012;Zhang et al., 2015;Miller and Raison, 2016).
Recently, our research group demonstrated that PGE 2 increases IL-1β production by a cyclic adenosine monophosphate (cAMP)-dependent mechanism in a model of sterile inflammation (Zoccal et al., 2016). Moreover, we identified that LTB 4 regulates this process by reducing intracellular cAMP levels, which in turn inhibits IL-1β release by innate immune cells (Zoccal et al., 2016). Considering that IL-1β and PGE 2 are potential triggers of depression (Müller et al., 2006;Goshen and Yirmiya, 2009), and that LTB 4 is able to repress the production of these mediators (Zoccal et al., 2016), we hypothesized that 5-LO deficiency would increase the susceptibility to depression. To investigate this hypothesis, we submitted Alox5 −/− (5-LO deficient) and 129sv mice to chronic unpredictable stress (CUS), an experimental model that induces depression-like behaviors such as decreased sucrose preference. Compared to 129sv mice at basal conditions, Alox5 −/− animals exhibited an intrinsically lower preference for sucrose and higher levels of IL-1β, IL-17, PGE 2 , adrenocorticotropic hormone (ACTH), and caspase-1 expression. In contrast, these mice present lower 11βHSD2 and GR expression, as well as reduced levels of corticosterone, 11-dehydrocorticosterone (11DHC), and circulating neutrophils. As expected, 129sv mice exposed to CUS displayed increased levels of splenic proinflammatory and stress-induced mediators, and signs of reduced sucrose preference, which could be associated with anhedonia (Willner, 1997). Intriguingly, most of the analyzed parameters remained at the same levels before and after CUS exposure in Alox5 −/− mice. These levels were comparable to those of CUS-129sv. Importantly, exogenous LTB 4 administration to CUS-Alox5 −/− mice decreased systemic corticosterone levels and caspase-1 expression without affecting sucrose preference.

Animals
Male mice (7-8-weeks-old) were used for the experiments. Animals lacking the 5-LO gene (Alox5 −/− ) were age-matched in all procedures with their genetic background, the wild-type 129sv mouse strain (Chen et al., 1994). The Alox5 −/− (129-Alox5 tm1Fun /J) and 129sv mice were obtained from The Jackson Laboratory (Bar Harbor, ME, USA) and were bred in the Faculdade de Ciências Farmacêuticas de Ribeirão Preto, Universidade de São Paulo (Ribeirão Preto, SP, Brazil). C57Bl/6 mice, obtained from the Faculdade de Ciências Farmacêuticas de Ribeirão Preto, Universidade de São Paulo, were used as controls to validate our methodology used in the 129sv mice. Mice were kept in 12 h light/dark cycles, with ad libitum access to water and food, except for the period of exposure to stressors (described in ''Chronic Unpredictable Stress and Sample Collection'' section). All procedures were performed according to the ethical standards established by the National Council for Control of Animal Experimentation (CONCEA) and approved by the Ethics Committee on Animal Use (CEUA) of the Campus of Ribeirão Preto (Protocol number: 14.1.290.53.5).

Chronic Unpredictable Stress and Sample Collection
C57BL/6, 129sv, and Alox5 −/− mice were subjected over a period of 14 days to the following daily stressors, which were applied randomly: an alteration in bedding (sawdust removal and replacement with wet sawdust having approximately 5 mm deep water) for 24 h; reversal of the light/dark cycle; 15 min forced swimming; 1 h restraint in a tube. The last stressor was always water and food deprivation for 20 h. The next day, animals underwent a sucrose preference test to evaluate depressive-like behavior (Willner, 1997). Non-stressed animals from each strain were used as controls and these animals were also subjected to 20 h without food and water before the sucrose test. In some experiments, mice exposed or not to CUS were also treated with specific drugs or vehicle, as described below.
At the end of the experiment, mice were anesthetized (75 mg/kg ketamine e 10 mg/kg xylazine) and the blood was immediately collected by cardiac puncture for the analysis of circulating leukocytes. All blood samples were collected at 12:00 pm and in the presence or absence of an anticoagulant and mice were weighed at day zero, and again at day fifteen. Next, the animals were euthanized by cervical dislocation. The spleen and adrenal glands, and the cerebellum, frontal cortex, and hippocampus were weighed after collection.

Sucrose Preference Test
129sv and Alox5 −/− mice were individually housed and deprived of food and water (the last stress stimulus) for 20 h. Next, a glass bottle with water and another with water containing 1% sucrose was placed in each cage. The bottles were initially weighed and consumption was monitored after 1 h. For this purpose, the glass bottles were weighed again to estimate the liquid ingestion. The sucrose preference was calculated as described previously (Guo et al., 2009) with the following formula: sucrose consumption/(sucrose consumption + water consumption).

LTB 4 Treatment in vivo
Alox5 −/− mice were intraperitoneally (i.p.) injected with LTB 4 (50 ng in 100 µL; Cayman, Ann Arbor, MI, USA) during the morning. Treatment initiated 1 day before CUS, followed by administration twice a day (12/12 h) during the 14 days of CUS exposure. LTB 4 was dissolved in ethanol and diluted in aqueous solution containing 0.45% NaCl, 0.25% dextrose, and 0.01% bovine serum albumin (BSA) followed by filter sterilization. The same volume of ethanol was added to the same aqueous solution to be used as the vehicle control. The LTB 4 dose was selected based on previous studies (Zoccal et al., 2016(Zoccal et al., , 2018.

Leukocyte Counts and Adrenocorticotropic Hormone (ACTH) Quantification
Blood samples were collected and immediately used for total and differential counts of circulating leukocytes as previously described (Pereira et al., 2013). In one specific set of experiments, blood was collected in tubes with EDTA (BD Microtainer, BD Biosciences, San Jose, CA, USA), centrifuged (400× g for 10 min at 4 • C), and the remaining plasma was stored at −20 • C. ACTH concentrations were determined with a chemiluminescent immunoassay (CLIA) technology (Liaison ACTH, 313221, DiaSorin, Sallugia, Piemonte, Italy) employing a Liaison Analyzer (DiaSorin, Sallugia, Piemonte, Italy), following the manufacturer's instructions. The sensitivity of the assay was <1,500 pg/mL.

Enzymatic Immunoassay (EIA) Measurements
Corticosterone levels were also determined using a specific enzyme immunoassay kit (Corticosterone ELISA kit, Enzo Life Science, New York, USA) according to the manufacturer's instructions. The reading was taken at 405-420 wavelength and the results were expressed in ng/mL.

Quantification of IL-1β and IL-17 by Enzyme-Linked Immunosorbent Assay
Spleens were homogenized (Mixer Homogenizer, Labortechnik, Wasserburg, Germany) and commercially available Enzyme-Linked Immunosorbent Assay (ELISA) kits (R&D Systems, Minneapolis, MN, USA) were used for quantification of IL-17 and IL-1β levels, according to the manufacturer's instructions. The sensitivity of the assay was <10 pg/mL.

Determination of T Lymphocyte Populations and Cellular Apoptosis by Flow Cytometry
Spleen cell suspensions were obtained by organ maceration following erythrocytes lysis with ACK lysis buffer (Invitrogen, Carlsbad, CA, USA). Total cell suspensions were adjusted to 1 × 10 6 cells/mL and flow cytometry was carried out as previously described (Secatto et al., 2012). The following fluorochrome-conjugated monoclonal antibodies were used: anti-CD3 (FITC); anti-CD4 (PE); and anti-CD8 (PerCP; BD Bioscience, San Jose, CA, USA). Samples were analyzed using a FACSCanto Flow Cytometer (BD Bioscience, San Jose, CA, USA). The percentage of CD3 + CD4 + and CD3 + CD8 + doublepositive cells were calculated using the FACSDiva software. Apoptosis was evaluated in spleen using an apoptosis detection kit based on annexin V-FITC/propidium iodide (PI) staining (BD Pharmingen, San Jose, CA, USA), according to the manufacturer's instructions. Briefly, cellular suspensions were washed twice with PBS, centrifuged at 400× g for 15 min, suspended in the kit buffer, and then incubated with annexin V in the absence of light for 15 min. PI was added to all samples immediately before data acquisition. A total of 10,000 events were acquired from all samples using the FACSCanto flow cytometer (BD Bioscience, San Jose, CA, USA). The results were evaluated using the FACSDiva software and expressed as the percentage of cells in the initial stage of apoptosis, i.e., positive for annexin V and negative for PI (Locachevic et al., 2015).

Quantitative Real Time PCR
Adrenal glands and spleens were collected, and total RNA was extracted using PureLink RNA Mini Kit (ThermoFisher Scientific, Waltham, MA, USA), according to the manufacturer's instructions. RNA concentration and purity were evaluated using a fluorimetric assay (Qubit, Invitrogen, Carlsbad, CA, USA). Complementary DNA (cDNA) was obtained using the High Capacity cDNA Reverse Transcription kit (Applied Biosytems, Foster City, CA, USA), according to the manufacturer's instructions. Total cDNA was amplified by real-time reverse transcription polymerase chain reaction (RT-qPCR) using TaqMan primers for the Hsd11b1, Hsd11b2, and Casp1 genes employing StepOne Plus (Applied Biosystems, Foster City, CA, USA). Actb was used as the endogenous control. The reactions were amplified under the following conditions: denaturation at 95 • C for 2 min, followed by 40 cycles of 95 • C for 2 min and 60 • C for 20 s. The expression of the genes Hsd11b1 (Mm01251104_m1, Applied Biosystems, Foster City, CA, USA), Hsd11b2 (Mm00476182_m1, Applied Biosystems, Foster City, CA, USA), and Casp1 (Mm.PT.58.13005595, IDT, San Diego, CA, USA) were normalized by housekeeping gene expression (Actb). The gene expression of 129sv mice was considered as the reference (129sv mice), being 2 −∆∆CT = 1. The primers were obtained commercially and are proprietary; thus, sequences are not available (TaqMan Gene Expression Assay, Applied Biosystems, Foster City, CA, USA).

Immunoblotting for Determination of the Glucocorticoid Receptor (GR)
The spleens and hippocampi from mice subjected or not to CUS were isolated and immediately immersed in RIPA buffer supplemented with protease inhibitors (Roche Diagnostics, Basel, Switzerland). After homogenization and centrifugation at 400× g for 10 min, the supernatant was used to measure GR expression. Equal amounts of protein (50 µg) from each sample were separated by polyacrylamide gel for electrophoresis (10% Tris-HCl) and subsequently transferred to a nitrocellulose membrane, as previously described (Serezani et al., 2005). The membranes were immunolabeled with an anti-GR monoclonal primary antibody (1:10,000; MCA1390, BioRad Antibodies, CA, USA), or anti-GAPDH antibody (1:5,000; Sigma Aldrich, St. Louis, Missouri, USA) as the endogenous control. Following membranes were incubated with the appropriate horseradish peroxidase (HRP)-conjugated secondary IgG (1:10,000; Cell Signaling, Danvers, MA, USA). ECL reagent (Amersham Biosciences, Little Chalfont, UK) was used to reveal by chemiluminescent reactions using a photo documenter (Alliance 4.7, UVITEC, Cambridge, England, UK). The band densities were determined using the ImageJ program 1 and normalized by the signal intensity emitted by the endogenous GAPDH control.

Enzymatic Activity of 11βHSD2
Activity of 11βHSD2 was evaluated as described by Kajantie et al. (2003) and Stewart et al. (1995) in samples of spleen, adrenal gland, frontal cortex and cerebellum from 129sv and Alox5 −/− mice, exposed or not to CUS. Briefly, all samples were homogenized (Mixer Homogenizer, Labortechnik, Wasserburg, Germany) in KCl buffer (0.154 mol/L; pH 7.6) and centrifuged at 1,000 g for 10 min 4 • C. Protein assay was performed on the supernatant using Coomassie blue staining, as described previously by Medeiros et al. (2008). Supernatants (0.50 g/L of spleen protein, 0.15 g/L of adrenal protein, 0.10 g/L of frontal cortex protein or 0.20 g/L of cerebellum protein) were incubated with a solution containing 1.0 mL of phosphate buffer (0.1 mol/L; pH 7.6), 0.8 nmol/L nicotinamide adenine dinucleotide, and 0.2 mol/L corticosterone in a shaking water bath at 37 • C. Spleen proteins were incubated for 30 min and the other tissues for 45 min. Enzymatic reactions were stopped with 1.0 mL of methanol. For extraction, 7 mL of dichloromethane was added to the samples and vortexed for 30 s. The upper phase was removed, 2 mL of water were added, and the tubes were mixed again. The lower phase (organic phase) was collected, dried and resuspended in 40 µL (spleen) or 25 µL (other tissues) of methanol. The enzymatic activity was evaluated by the conversion of corticosterone into 11DHC, which were measured by LC-MS/MS (Peti et al., 2018).

Statistical Analyses
Data were analyzed using SPSS V. 20 (IBM, USA). After checking for normality and homoscedasticity, the results were analyzed using Student's t-test, two-or one-way analysis of variance (ANOVA), when appropriated. Post hoc comparisons were performed with the Newman-Keuls test. Differences were considered significant when p < 0.05.

Alox5 −/− Mice Show an Intrinsic HPA Activation and Differential Dynamics of Systemic and Hippocampal Corticosterone
To determine the impact of 5-LO deficiency on stress-induced behavior and immune response, we began by comparing the amount of circulating corticosterone in two distinct mouse strains, C57Bl/6 and 129sv. The former strain has been widely characterized in studies addressing CUS-induced depression-like behavior (Pothion et al., 2004). Both mouse strains exhibited similar levels of serum corticosterone before CUS exposition ( Figure 1A). Following 14 days of CUS exposure, serum levels of corticosterone increased significantly in both strains (F (1,22) = 42.06, p < 0.001), but corticosterone levels remained comparable between them ( Figure 1A). We observed similar blood neutrophil counts between both mouse strains at basal conditions, which were equally reduced after CUS exposition (F (1,23) = 7.13, p < 0.02; Supplementary Figure S1B). Mononuclear cells were also equally represented in the blood of both strains at basal conditions, but these cells were significantly reduced in the blood of CUS-129sv compared to CUS-C57Bl/6 (interaction CUS vs. strain, F (1,19) = 6.83, p = 0.017; Supplementary Figure S1A). Considering that 129sv mice submitted to CUS presented a similar profile in these parameters to that of C57Bl/6 mouse strain, we were confident to use the 129sv strain as controls of experiments performed with Alox5 −/− mice.
Deficiency of 5-LO abrogates the production of LTs and cysteinyl-LTs. Considering that HPA axis activation and corticosterone production are important components of the stress response, we sought to determine the impact of 5-LO deficiency on CUS-induced ACTH release in plasma samples, corticosterone production, and its metabolization into 11DHC in serum. At basal conditions, Alox5 −/− exhibited significant lower amounts of circulating corticosterone and 11DHC compared to 129sv mice (Corticosterone, F (1,20) = 15.70, p < 0.001; 11DHC, F (1,20) = 1239.0, p < 0.001; Figures 1B,C).We also found higher ACTH levels in Alox5 −/− mice (CUS vs. strain interaction; ACTH, F (1,15) = 7.98, p = 0.013; Figure 1D). Compared to non-stressed 129sv controls, CUS-129sv mice produced 97% more circulating corticosterone and approximately 32% less circulating ACTH (Figures 1B,D). CUS-Alox5 −/− mice produced 87% more circulating corticosterone and had a reduction of approximately 64% in ACTH levels (Figures 1B,D). Together these results suggest that 5-LO deficiency results in lower corticosterone concentrations in the circulation, which triggers the activation of the HPA axis, reflected by high levels of plasma ACTH in Alox5 −/− mice at basal conditions. Furthermore, these data suggest similar hormonal profiles in the blood of Alox5 −/− mice at basal conditions and 129sv exposed to CUS.
Because depression affects primarily the brain, we investigated some aspects of stress response in the hippocampus of 129sv and Alox5 −/− mice. Both strains showed similar amounts of corticosterone ( Figure 1E) and 11DHC ( Figure 1F) at basal conditions. After CUS exposition, corticosterone levels equally increased in the hippocampus in both strains, but no changes were detected in 11DHC levels. Corticosterone levels are self-regulated by activation of GR. Thus, we hypothesized that the discrepancies between serum and hippocampus corticosterone levels in 129sv and Alox5 −/− mice could be related to lower GR expression, with consequent deregulation of the negative feedback loop in the HPA axis. We found that cells isolated from the hippocampus of non-stressed Alox5 −/− mice showed a trend to reduced GR expression (p = 0.08; Figure 1G). Importantly, hippocampal GR expression was significantly reduced in CUS-129sv, but not CUS-Alox5 −/− mice.

Alox5 −/− Mice Have a Similar Profile of Stress-Induced Mediators to That of CUS-129sv Mice-Brain Structures Panorama
Contrasting data between blood and hippocampal corticosterone levels prompted us to perform a detailed investigation of different brain structures. Compared to 129sv mice, Alox5 −/− animals showed a significant increase in corticosterone production in the frontal cortex at basal conditions (Figure 2A). After CUS FIGURE 1 | Alox5 −/− mice exhibit a natural hypothalamic-pituitary-adrenal (HPA) axis activation and produce different corticosterone amounts in the serum and hippocampus. (A) The C57Bl/6 and 129sv mice were subjected or not to chronic unpredictable stress (CUS) for 14 days. Animals were euthanized on the 15th day and blood was collected for corticosterone quantification. (B-D) Alox5 −/− and 129sv mice were subjected or not to CUS for 14 days. Animals were euthanized on the 15th days and blood was collected for quantification of (B) corticosterone, (C) 11-dehydrocorticosterone (11DHC) and (D) adrenocorticotropic hormone (ACTH).
The 11βHSD1 and 11βHSD2 enzymes are responsible for the metabolism of corticosterone, with 11βHSD1 promoting its production, whereas 11βHSD2 converts it into the inactive metabolite 11DHC. To investigate the potential molecular mechanism regulating the balance between corticosterone and 11DHC, we assessed the enzymatic activity of 11βHSD2 in the frontal cortex and cerebellum. Compared to 129sv mice, 11βHSD2 activity increased in the frontal cortex of Alox5 −/− and CUS-129sv mice (F (1,17) = 8.237, p = 0.0106; F (1,17) = 12.39, p = 0.0026; Figure 2E) and cerebellum of CUS-129sv (F (1,17) = 85.16, p < 0.0001; Figure 2F). Despite the increased 11βHSD2 activity in the frontal cortex from CUS-129sv and in Alox5 −/− mice, the reduced 11DHC levels suggest that corticosterone is produced at a higher rate than consumed, as a result of a higher 11βHSD1 activity. It is also FIGURE 2 | Alox5 −/− and CUS-129sv mice exhibit similar hormonal profile in different brain structures. 129sv and Alox5 −/− mice were subjected or not to CUS for 14 days. The animals were euthanized on the 15th day and the brains collected to dissect the frontal cortex and cerebellum. Quantification of (A,B) corticosterone (C,D) 11DHC and (E,F) 11β-hydroxysteroid dehydrogenase (11βHSD2) enzymatic activity in the frontal cortex and cerebellum, respectively. Data are presented as the means ± SEM of one experiment (n = 6-11 mice/genotype/group). Two-way ANOVA followed by Newman-Keuls Multiple Comparison Test, p < 0.05. * 129sv vs. CUS-129sv; # 129sv vs. Alox5 −/− ; & CUS-129sv vs. CUS-C57Bl/6 or CUS-Alox5 −/− . possible that corticosterone is catabolized by a different pathway that converts it into a metabolite other than 11DHC.

Alox5 −/− Mice Exhibit a Natural Low Sucrose Preference at Basal Conditions-Behavior Panorama
We evaluated Alox5 −/− and 129sv strains, submitted or not to CUS, for sucrose preference, to determine the involvement of 5-LO deficiency on that behavior. The test assesses the reduction in sugar solution intake and sucrose preference in relation to a control group (Guo et al., 2009). Exposure to CUS decreased sucrose solution, but not total liquid intake by CUS-129sv mice compared to 129sv animals in basal conditions (CUS, F (1,36) = 8.10, p = 0.007, CUS vs. strain, F (1,36) = 4.68, p = 0.03; Figure 3A). This was reflected by lower sucrose preference (F (3,36) = 5.11, p = 0.005; Figure 3B), indicating that 129sv mice, which express 5-LO, develop decreased sucrose preference when exposed to CUS. However, compared to non-stressed 129sv mice, Alox5 −/− animals at basal conditions already exhibited lower sucrose preference, which remained unchanged after 14 days of CUS exposure (Figure 3B). These results demonstrate that the Alox5 −/− mice present a natural lower sucrose preference, which is comparable to 129sv mice exposed to CUS. Furthermore, that behavior was not altered in Alox5 −/− mice. Therefore, impaired LTB 4 production could predispose these animals to reach a maximal decreased sucrose preference response that remains unchanged even under conditions of chronic stress. We also weighed the whole body of 129sv and Alox5 −/− before and after 14 days of CUS ( Figure 3C).CUS exposure significantly reduced the body weight in both mouse strains (F (1,20) = 8.41, p < 0.01). Supporting our previous findings, these results confirm a significant impact of 5-LO deficiency in a natural depressive-like phenotype of Alox5 −/− mice that remains mostly unaltered after stressful stimuli.
FIGURE 3 | 5-lipoxygenase (5-LO) deficiency results in decreased sucrose preference at basal conditions. 129sv and Alox5 −/− mice were exposed or not to CUS for a period of 14 days. Animals were deprived of water and food for 20 h on the 14th day. (A,B) On the 15th day, mice were evaluated for (A) liquid intake and (B) sucrose preference. (C) 129sv and Alox5 −/− mice were weighed on day 0, and exposed or not to CUS for a period of 14 days. Animals were weighed again on the 15th day, and the weight changes were expressed as a percentage before CUS. Data are presented as the means ± SEM and depict a representative experiment out of three independent experiments (n = 6-11 mice/genotype/group/experiment). Two-way ANOVA followed by Newman-Keuls Multiple Comparison Test, p < 0.05. * 129sv vs. CUS-129sv; # 129sv vs. Alox5 −/− . Frontiers in Behavioral Neuroscience | www.frontiersin.org FIGURE 4 | 5-LO regulates the production of inflammatory mediators before and after CUS exposition. 129sv and Alox5 −/− mice were exposed or not to CUS for 14 days. Animals were euthanized on the 15th days and spleens were collected for quantification of (A) IL-1β, (B) IL-17, (C) PGE 2 and (D) leukotrienes B 4 (LTB 4 ). Data are presented as means ± SEM of one representative out of two independent experiments (n = 6 mice/genotype/group/experiment). Two-way ANOVA followed by Newman-Keuls Multiple Comparison Test, p < 0.05. * 129sv or Alox5 −/− vs. CUS-129sv or CUS-Alox5 −/− ; # 129sv vs. Alox5 −/− ; & CUS-129sv vs. CUS-Alox5 −/− . ND, not determined.
Together, our results indicate that 5-LO derived metabolites regulate the apoptotic cellular death in the spleen, which associates with increased caspase-1 expression and decreased 11βHSD2 activity at basal conditions, that remains at the same levels after chronic stress.

LTB 4 Regulates Levels of Systemic Corticosterone and Caspase-1 Expression
At this point of our study, we have demonstrated that Alox5 −/− mice exhibit depressive-like phenotype at basal conditions. Excluding PGE 2 production, CUS exposure does not seem to cause a significant impact on this profile, which is very similar to that of CUS-129sv mice. Considering that Alox5 −/− animals lack all 5-LO-derived metabolites and LTB 4 is the major  product of this pathway (Samuelsson, 2000), we performed a pharmacological treatment with this lipid mediator to evaluate its specific involvement on the altered biological and behavioral responses due to 5-LO deficiency (Supplementary Figure S2). Interestingly, LTB 4 reduced serum corticosterone levels of CUS-Alox5 −/− mice compared to both Alox5 −/− and CUS-Alox5 −/− mice at basal conditions treated with vehicle ( Figure 7A). In contrast, all groups exhibited similar sucrose preference ( Figure 7B). Together, these results suggest that, at least in this model, corticosterone does not mediate the decreased sucrose preference. We have shown previously that LTB 4 inhibits IL-1β release during sterile inflammation (Zoccal et al., 2016(Zoccal et al., , 2018. Therefore, we evaluated whether this change was associated with the levels of caspase-1 expression. Compared to vehicle-injected CUS-Alox5 −/− mice, LTB 4 treatment significantly decreased caspase-1 expression in the adrenal glands (F (2,8) = 10.52, p < 0.06; Figure 7C) and in the spleen (F (2,12) = 5.91, p = 0.016; Figure 7D). These results could help to explain the high levels of IL-1β production of Alox5 −/− mice ( Figure 4A). Taken together, our results indicate that, besides inhibiting intracellular cAMP production (Zoccal et al., 2016(Zoccal et al., , 2018, LTB 4 also inhibits caspase-1 expression to suppress IL-1β release. Moreover, these data demonstrate that LTB 4 also regulates the levels of systemic corticosterone production after CUS exposure. , encoded by the Alox5 gene, catalyzes the formation of (LTB 4 ) which is important for the regulation of the HPA-axis and consequently corticosterone levels. This study showed that conditional Alox5 −/− mice present reduced sucrose preference, suggesting reduced reward sensitivity. Furthermore, caspase-1, IL-1β, IL-17 and PGE 2 increased in these animals. In frontal cortex corticosterone FIGURE 7 | Exogenous LTB 4 decreases corticosterone production and caspase-1 expression, but not decreased sucrose preference in Alox5 −/− mice. Animals were subjected or not to CUS over a period of 14 days. CUS-Alox5 −/− mice were treated twice a day (12 h/12 h) with LTB 4 (50 ng/animal/100 µL i.p.) or vehicle (100 µL i.p.). (A) Blood was collected on the 15th day of quantification of circulating corticosterone. (B) Before euthanasia on the 15th day, mice were evaluated for sucrose preference. (C) Adrenal glands and (D) spleens were removed for evaluation of the transcriptional caspase-1 expression. In (C,D), 129sv mice at basal conditions were used as a reference (continuous line), and Alox5 −/− mice at basal conditions were used as controls. Data are presented as means ± SEM of one experiment with n = 6-8 mice/genotype/group. Two-way ANOVA followed by Newman-Keuls Multiple Comparison Test, p < 0.05. * Alox5 −/− vs. CUS-Alox5 −/− + Vehicle or CUS-Alox5 −/− + LTB 4 ; & CUS-Alox5 −/− + Vehicle vs. CUS-Alox5 −/− + LTB 4 . also increased, although its systemic levels decreased. The observed alterations were in general similar to those observed in the CUS-induced changes in control animals. In this case, however, systemic corticosterone levels increased and, notably, LTB 4 levels decreased. LTB 4 administration restored caspase-1 levels and systemic corticosterone levels, but not reduced sucrose preference.

DISCUSSION
In the past few years, several reports have provided evidence for the direct relationship between depression and inflammation. However, the role of 5-LO-derived metabolites in depression onset and progression is unclear (Uz et al., 2008;Athayde et al., 2009;Chu and Praticò, 2009;Joshi and Praticò, 2013;Joshi et al., 2015). LTB 4 is the main product of 5-LO activation. CUS exposure of 129sv mice, however, had a modest impact over LTB 4 production in the spleen (p = 0.054). In line with these findings, reduced (CUS-129sv) or absence (Alox5 −/− ) LTB 4 production influenced several parameters associated with stressinduced behavior, although we cannot exclude the influence of other 5-LO related metabolites. Alox5 −/− animals at basal conditions already exhibited lower interest for sucrose. We also detected, using the tail suspension test (TST), that Alox5 −/− mice exhibited a natural depressive-like behavior, assessed through the immobility induced by their inability or reluctance to maintain effort rather than a generalized hypoactivity. CUS exposition, however, did not change that behavior. Besides that, we observed that CUS exposure did not cause a significant impact on the immobility time of 129sv, indicating that 14 days of CUS exposure were not enough to change this behavior, contrasting with sucrose preference (data not shown). Furthermore, exposure to CUS did not impact any of the evaluated behavioral parameters in Alox5 −/− mice, which were, however, similar to those observed in CUS-129sv animals. This finding shows that 5-LO-deficiency results in decreased sucrose preference that is not modified by chronic stress, possibly indicating a ceiling effect. One limitation of our study is that 129sv and Alox5 −/− mice were not littermates, not sharing, therefore, the same intra-uterine and preweaning environment. This limitation is relevant because the maternal behavior could have been affected by 5-LO deletion and, consequently, affected the phenotype of the adult offspring (Gingrich and Hen, 2000). Although the pharmacological results, discussed above, suggest that even this possibility exists, the involvement of LTB 4 in the observed behavioral changes is likely. Future studies are necessary to confirm whether the phenotypes observed here were replicated when 129sv and Alox5 −/− were born and weaned by the same mother.
Of importance, Alox5 −/− mice at basal conditions displayed increased levels of circulating ACTH, indicating that 5-LO deficiency triggers the HPA axis activation even in the absence of stressful stimuli. However, systemic levels of corticosterone were reduced in Alox5 −/− mice at basal conditions. They increased after exposure to CUS but did not reach the same levels of CUS-129sv mice. Therefore, our data corroborate the proposal that 5-LO and related metabolites regulate the activity of HPA axis (Joshi and Praticò, 2013). We hypothesized that 11βHSD activity would be associated with the reduced levels of circulating corticosterone. To test this, we investigated the expression and/or activity of 11βHSD isoforms. The regulation of 11βHSD1 and 11βHSD2 expression is a complex process that directly controls corticosterone levels. Interestingly, IL-1β FIGURE 8 | Alox5 −/− mice display a depressed-like behavior at basal conditions comparable to that of 129sv animals exposed to CUS. In general, corticosterone levels and 11βHSD2 activity increased and 11DHC levels decreased in the brains of Alox5 −/− and CUS-129sv mice. These animals also exhibited increased caspase-1 expression in the adrenal gland and spleen associated with increased levels of splenic IL-1β. In addition, 5-LO deficiency at basal conditions or CUS exposure of 129sv mice decreased 11βHSD2 mRNA expression and/or activity in the adrenal gland and spleen, while increasing IL-17 and PGE 2 or decreasing LTB 4 production in the spleen. However, 5-LO deficiency culminates in elevated basal activation of the HPA axis, reflected by high levels of circulating ACTH but reduced levels of systemic corticosterone, possibly due to the impaired expression of the ACTH receptor (MC2R) in the adrenal glands (systemic panorama-left panel). In contrast, CUS-129sv mice exhibited reduced circulating levels of ACTH and increased corticosterone (systemic panorama-right panel) and decreased GR expression in the hippocampus (brain panorama-right panel).
increases the expression of 11βHSD1 (Tetsuka et al., 1999), while reducing the expression and activity of 11βHSD2 (Chisaka et al., 2005). In line with these facts, our results show that CUS-129sv mice presented increased IL-1β levels and 11βHSD1 expression in the spleen, concomitantly with reduced 11βHSD2 expression. The reduced activity of 11βHSD2 in their spleens or adrenal glands, however, suggest that 5-LO and related metabolites engage distinct molecular mechanisms to regulate levels of systemic corticosterone.
Contrasting with the spleen and adrenal glands, the frontal cortex of Alox5 −/− and CUS-129sv mice exhibited increased 11βHSD2 activity. These discrepancies suggest that 5-LO deficiency or CUS exposition results in organ-dependent changes. Future studies are needed to elucidate the mechanisms underlying such differences.
The basal inflammatory state of an individual is highly relevant in the context depression (Maes et al., 1997). We demonstrated that Alox5 −/− mice at basal conditions already exhibit increased levels of splenic IL-1β and IL-17, which are associated with behavioral changes. They also develop exacerbated inflammatory responses and become highly susceptible to a model of sterile inflammation (Zoccal et al., 2016), suggesting that the molecular mechanisms underlying the behavioral modifications of Alox5 −/− mice at basal conditions are directly involved with a deregulated basal inflammatory profile. Thus, non-stressed Alox5 −/− animals could already release cytokines at high levels (i.e., the ceiling effect), justifying why stress did not cause a further increase of these mediators.
Strikingly, we found that CUS exposition results in increased PGE 2 production, whereby CUS-Alox5 −/− mice exhibited even higher levels compared to CUS-129sv. We have demonstrated that PGE 2 boosts IL-1β production by innate immune cells (Zoccal et al., 2016(Zoccal et al., , 2018. The release of this cytokine is conditioned to the activation of inflammatory caspases (Friedlander et al., 1996;Rano et al., 1997). Higher levels of PGE 2 were associated with increased caspase-1 expression in basal conditions of Alox5 −/− and CUS-129sv mice, suggesting that, beyond PGE 2 signaling, LTB 4 also represses caspase-1 expression to limit IL-1β release. Paugh et al. (2015) demonstrated that higher caspase-1 expression augments GR cleavage, resulting in elevated glucocorticoid resistance. GR regulation is an important checkpoint during corticosterone production. Of note, GR expression was significantly reduced in the hippocampus of CUS-129sv, suggesting that levels of systemic corticosterone increase due a disruption on negative feedback, which was not observed in Alox5 −/− . Despite the increased circulating levels of ACTH in Alox5 −/− mice, they exhibited reduced corticosterone levels. These levels increased after CUS but remained significantly lower compared to CUS-129sv mice. Thus, 5-LO and related metabolites might regulate the systemic levels of corticosterone by promoting the expression and signaling of the ACTH receptor (MC2R) in the adrenal glands. However, future studies are needed to test this hypothesis.
Compared to Alox5 −/− mice at basal conditions, CUS-Alox5 −/− show increased levels of systemic corticosterone. Treatment with LTB 4 reduced circulating corticosterone but did not cause any alteration on sucrose preference. Considering that Alox5 −/− mice are unable to produce other molecules beyond LTB 4 , we cannot discard the involvement and synergy between LTC 4 , D 4 or E 4 in sucrose preference behavior. A recent study demonstrated that the knockdown if the cysteinyl-LT receptor 1 in the hippocampus reverses the depression-like behavior of mice submitted to CMS (Yu et al., 2016). LTB 4 treatment was unable to reduce the decreased sucrose preference, but regulated caspase-1 expression, IL-1β release, and corticosterone production. These findings suggest that, while LTB 4 controls the production of mediators involved in the inflammatory response, decreased sucrose preference could be directly regulated by cysteinyl-LTs. Further investigation will reveal the real contribution of these cysteinyl-LTs and LTB 4 to the onset and progression of depression. Patients with major depressive disorders have lower levels of eicosanoid precursors, the polyunsaturated fatty acids (PUFA), which is associated with the severity of depressive symptoms (Adams et al., 1996). Antidepressant effects in these patients might involve eicosanoid cascades such as PGE 2 formation (Su, 2009). Drugs such as imipramine, desipramine, and clomipramine reduce systemic levels of corticosterone in rats submitted to force swimming stress test, which prolonged swimming time (Yamada et al., 1987). Similar to LTB 4 treatment, several antidepressants reduced caspase-1 expression and IL-1β production (Alcocer-Gómez et al., 2017). While LTB 4 treatment did not affect sucrose preference, decreased inflammation induced by the treatment with LTB 4 or antidepressants might account for changes in other behavioral parameters. A limitation of this study is that there was no group treated with antidepressant. Therefore, future studies comparing the effects of LTB 4 and antidepressants in 129sv and Alox5 −/− mice are needed to address this problem.
A previous work demonstrated that exogenous administration of LTB 4 activates the HPA axis via BLT1 receptor, increasing plasma levels of ACTH and corticosterone while downregulating the inflammatory response during asthma (Zhang et al., 2010). However, our study is the first to demonstrate that 5-LO and related metabolites promote differential regulation of the systemic levels of ACTH and corticosterone, impacting stress and inflammation. This is highly relevant to explain the controversial findings of the role of LTB 4 in infectious diseases. It is well recognized that LTB 4 increases phagocytosis of microorganisms in vitro (Serezani et al., 2005;Secatto et al., 2014;Prado et al., 2017), as does dexamethasone (van der Goes et al., 2000). However, the results regarding LTB 4 function in vivo are conflicting. For example, different outcomes have been described for experimental tuberculosis using LTB 4 pharmacological inhibition or gene deletion. Indeed, treatment with MK886 inhibits LTB 4 production and increases bacterial burden and mortality of mice infected with Mycobacterium tuberculosis (Peres et al., 2007). In contrast, Alox5 −/− mice are more resistant to M. tuberculosis infection, have lower bacterial burden, and survive longer than 129sv animals (Bafica et al., 2005). These opposing results generate a paradox regarding the importance of LTB 4 in the control of long-lived bacterial infections such as M. tuberculosis. Based on our present data, we suggest that the low levels of corticosterone observed in Alox5 −/− mice favors the host immune response by promoting a quick pathogen elimination. Although major depression is a polygenic condition and depends on the interaction with environmental factors (Kendler et al., 1995), the results indicate that a decrease in 5-LO expression could be associated with this disorder.
In summary (Figure 8; Supplementary Table S1), we demonstrated that CUS results in significant modifications in 129sv mice, increasing corticosterone levels and reducing GR expression in hippocampus. In addition, the HPA axis is activated, resulting in increased circulating corticosterone and a series of other organ responses, such as increased caspase-1 expression in the spleen and adrenal glands; increased levels of splenic 11βHSD1 expression and IL-1β and PGE 2 production; decreased levels of splenic LTB 4 and 11βHSD2 expression. Basically, the same phenomena occur in Alox5 −/− mice at basal conditions. 5-LO and related metabolites control caspase-1 expression, IL-1β, and PGE 2 release. This inflammatory environment possibly increases HPA activation. The adrenal glands, however, seem to be unable to produce enough corticosterone, which results in lower circulating amounts of this hormone, even after CUS exposition.
In conclusion, this study provides important insights into the role of 5-LO and related metabolites in the control of stress-induced mediators and inflammation. suggesting that Alox5 −/− mice could be a novel model for the understanding of the molecular mechanisms responsible for decreased sensitivity to rewards.

ETHICS STATEMENT
This study was carried out in accordance with the recommendations of National Council for Control of Animal Experimentation (CONCEA). The protocol was approved by the Ethics Committee on Animal Use (CEUA) of the Campus of Ribeirão Preto (Protocol number: 14.1.290.53.5).

AUTHOR CONTRIBUTIONS
GL designed the research, analyzed the data and prepared the figures. GL and MP executed the main experimental procedures and wrote the article. KZ, PP, MB, and AP assisted with laboratory procedures. MF assisted with initial CUS protocol. LF, CS, EDB and FG analyzed the data and discussed the hypothesis. FG performed statistical analysis. LF conceived and supervised the project, designed the experiments, obtained the funding grants, helped with the data interpretation, wrote and discussed the manuscript and acted as scientific advisor.

FUNDING
This work was supported by the São Paulo Research Foundation (Fundação de Amparo à Pesquisa do Estado de São Paulo, FAPESP; grant #2014/07125-6 to LF; EMU grant #2015/00658-1 to LF), Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES-Finance Code 001), and from Conselho Nacional de Pesquisa (CNPq). The funders had no role in the study design, data collection and analysis, decision to publish, or preparation of the article.