Chronic Unpredictable Stress Alters Brain Tryptophan Metabolism and Impairs Working Memory in Mice without Causing Depression-Like Behaviour

Chronic stress is a well-known risk factor in major depressive disorder and disrupts the kynurenine and serotonin pathways of tryptophan metabolism. Here, we characterize the temporal central and peripheral changes in tryptophan metabolism and concomitant depressive-like behavioural phenotype induced during the progression of chronic unpredictable stress (CUS). Mice were exposed to 0, 10, 20, or 30 days of CUS followed by a panel of behavioural assays to determine depressive-like phenotypes. Immediately after behavioural testing, plasma and brain tissue were collected for metabolic analysis. While anhedonia-like and anxiety-like behaviours were unaffected by stress, nesting behaviour and cognitive deficits became apparent in response to CUS exposure. While CUS caused a transient reduction in circulating quinolinic acid, no other tryptophan metabolites significantly changed in response to CUS. In the brain, tryptophan, kynurenine, picolinic acid, and 5-hydroxyindoleacetic acid concentrations were significantly elevated in CUS-exposed mice compared with non-stress control animals, while kynurenic acid, xanthurenic acid, and serotonin decreased in CUS-exposed mice. Metabolic turnover of serotonin to the major metabolite 5-hydroxyindoleacetic acid was markedly increased in response to CUS. These results suggest that CUS impairs hippocampal-dependent working memory and enhances nascent nesting behaviour in C57BL/6J male mice, and these behaviours are associated with increased brain kynurenine pathway metabolism leading to accumulation of picolinic acid and a significant reduction in serotonin levels.


Introduction
Depression is a major public health concern affecting approximately 300 million people globally [1]. Chronic stress is a well-characterized risk factor in the development of depression [2]. Depressed individuals have elevated levels of cortisol in their saliva and plasma, as well as dysregulated feedback inhibition of the HPA axis [3,4]. Further, antidepressant treatment reduces HPA activity in animals and depressed humans [5]. Chronic unpredictable stress (CUS), social defeat, restraint, and other stressors are common pre-clinical models used in the investigation of the pathogenesis of depression [6].
A key component of the physiological response to chronic stress is the interaction with the immune system [7][8][9][10]. Pro-inflammatory markers in the central nervous system (CNS) have been highly implicated in the pathogenesis of depression [11][12][13][14]. Mounting evidence suggests that chronic stress up-regulates the expression of pro-inflammatory cytokines in the CNS through activation of resident microglia [15][16][17]. Il-1β, in particular, has been shown to be strongly up-regulated in response to chronic mild stress in mice, while genetic or pharmacologic blockade of Il-1β receptor ameliorates the stress-induced depressive-like behavioural response [18,19]. Over the past two decades, alterations in the kynurenine pathway of tryptophan metabolism have been implicated in the pathogenesis of inflammation-induced depression, and preclinical models have largely recapitulated these findings [12,20]. However, relatively less is known pertaining to the potential pathogenic role of tryptophan metabolism along the kynurenine pathway. Neurology  Tryptophan is metabolized to kynurenine through the indolamine enzymes (indolamine-2,3-dioxygenase and tryptophan dioxygenase; IDO1, IDO2, and TDO) ( Figure 1). Downstream kynurenine metabolism then diverges into two branches to produce the neuroactive metabolites kynurenic acid (KA) and quinolinic acid (QA). Under baseline conditions, the majority of kynurenine is converted to KA through the enzyme kynurenic acid transferase (KAT). However, this balance is shifted during an inflammatory response, resulting in up-regulated expression of IDO and kynurenine monooxygenase (KMO) in microglial cells. KMO initiates the oxidative branch of the kynurenine metabolism pathway by metabolizing kynurenine into 3-hydroxykynurenine (3-HK), the precursor to QA. Elevated levels of QA in the brain have been associated with depressive symptoms in humans and inflammation results in a robust increase in kynurenine metabolism that mediates depressive-like behaviours and cognitive impairment [21][22][23]. This metabolic change is characterized by a shift towards oxidative kynurenine metabolism, particularly in the hippocampus, while serotonin levels are largely unaffected [24,25]. Kynurenine metabolism via the cortisol-dependent enzyme TDO has been implicated in mediating KP metabolism during stress, but experimental data are limited [26]. Gibney et al. found that the non-selective TDO inhibitor attenuated depressive-like behaviour in rats exposed to repeated restraint stress, while Dostal et al. found that acute restraint stress in mice upregulated TDO in an IDO1-dependent manner to increased kynurenine levels, but no phenotypic measures were reported in the latter study [27,28]. In this study, central and peripheral kynurenine pathway metabolites and depressive-like behavioural changes were measured in C57Bl/6J mice following varying durations of CUS exposure. Here, we report that CUS exposure failed to precipitate a depressive-like phenotype, but nascent nesting behaviour was increased and working memory impaired by CUS exposure. These behavioural phenotypes were paralleled by increased levels of tryptophan and certain downstream kynurenine pathway metabolites (kynurenine and picolinic acid) in the brain, while central kynurenic acid and xanthurenic acid and peripheral quinolinic acid were reduced. Furthermore, brain serotonin concentration levels were significantly decreased following prolonged stress exposure, while 5-HIAA was greatly increased. These results suggest that the tryptophan in the brain is shunted towards the kynurenine metabolism pathway and coupled with increased serotonin turnover, brain serotonin levels are markedly reduced. Together, these observations suggest that dysregulated tryptophan metabolism pathways are a possible mechanistic link between chronic stress and depressive-like phenotypes.

I Mice
8-10-week-old male C57BL/6J mice (The Jackson Laboratory; Bar Harbor, ME; stock # 000664) were housed in groups of 2-4 mice/cage and maintained on a 12:12h reverse light cycle (lights off at 11:00). Food and water were available ad libitum. Body weights were recorded throughout the experiment to track physiological responses to stress. All animal care and use were carried out in accord with the Guide for the Care and Use of Laboratory Animals, 8th edition (NRC) and approved by the Institutional Animal Care and Use Committee at The University of Texas Health San Antonio. Mice were exposed to 0, 10, 20, or 30 days of chronic unpredictable stress (n=8/group). On the last day of stress exposure, a battery of behavioural assays was performed. Behavioural analysis of mice exposed to 0 days of CUS was performed in conjunction with Groups 1-3. Immediately following behavioural assessments, mice were euthanized via carbon-dioxide asphyxiation and exsanguination. Blood was collected for plasma analysis. Mice were then perfused with heparinized saline. Brain tissue and spleens were collected and snapfrozen in liquid nitrogen. All tissue and plasma were stored at -80°C. Brain tissue was homogenized with mortar and pestle on dry ice to be divided into equal samples for metabolite and protein analysis (described below; n=4-6/group).

II Chronic Unpredictable Stress Paradigm
All procedures were initiated during the dark phase of the light cycle, unless otherwise noted, and were carried out in the room the CUSexposed mice were housed in. Mice exposed to 0 days of stress were housed in a separate room on the same light cycle. The stressors were applied randomly to maximize unpredictability; each stressor was only repeated once every 10 days (Table 1) and at varying times during the dark phase. The mice were placed in their home cages on an orbital shaker set to 70 RPM. The orbital shaker was set on a timer for one hour on/off intervals for 24 hours.
ii Cold Swim 4-liter cylindrical clear containers were filled with tap water and ice to maintain temperatures of 17-19°C. Mice floated or swam in the cold water for 10 minutes. After 10 minutes, mice were removed from the cold water and dabbed with paper towels.
iii Restraint 50-ml Falcon conical tubes were modified by removing the tip to allow airflow and 3mm holes were drilled in the caps to allow access to their tails. Mice were held in these tubes with the cap screwed on behind them for 2 hours.

iv Restraint with a Tail Pinch
Mice were held for 2 hours in 50-ml tubes as described. In the last 2 minutes of restraint, plastic clothespins were clipped onto the base of their tails, exposed through the caps.

v Predator Odor
The gauze was soaked with 2ml of coyote urine (Coyote Pee Brand Coyote Urine; https://www.predatorpeestore.com/coyote-urine-16-oz-33-day-dispenser-combo_moreinfo.html) and secured to the inside of the micro-isolator tops of the mouse home cages for 12 hours. After 12 hours, the mice were placed into fresh cages with new micro-isolator tops and the gauze was disposed of outside the animal housing facility.

vii Wet Bedding
Approximately 300-400ml of water was added to home cages to sufficiently soak bedding with no standing water. After 24 hours, mice were moved to cages with fresh bedding.

viii Strobe Light
A strobe light was pointed at the home cages of the mice for 12 hours during the dark phase of the light cycle.

ix Warm Cage
The micro-isolator top of the home cages was removed and a 75watt heat bulb was pointed over the cages so that the temperature was maintained at approximately 35° (underneath bulb) to 26° (periphery of the cage).
The heat bulbs were turned on for one-hour on-off intervals for 12 hours.
x Reverse Light Cycle For 12 hours during the dark phase of the light cycle, 120 lux bulbs were pointed at the home cages of the mice.

xi Modified Social Defeat
Experimental mice were placed in the home cage of 2 adult male CD1 mice (Charles River; strain code 022) for 5 minutes. One hour later, experimental mice were placed in the home cage of 2 different CD1 male mice for 5 minutes.

xii Isolation
All experimental mice are singly housed for 48 hours before returning to their cage mates.

III Behaviour Analysis i Saccharin Preference
Three days prior to CUS exposure, mice were trained to differentiate water and 0.04% saccharin using a two-bottle choice testing paradigm. Bottles were placed above the food hopper and their location was alternated daily to control for a potential place preference. On the last day of CUS, the two bottles were returned to the cages. Saccharin preference was calculated as (saccharin intake)/(water + saccharin intake)*100 and interpreted as a measure of anhedonia-like behaviour.
ii Nesting Nesting behaviour was assessed in the home cage of individually housed mice following the zone clearance method [29]. Nestlet pieces were divided into 6 equal pieces and placed in each corner of the home cage ("zones"; Figure 2B). Mice instinctively gather up the nestlet pieces to build one larger nest. The number of nestlet pieces that have been cleared from their respective zones is counted every 6 minutes for a maximum score of 5 zones cleared in 42 minutes ( Figure 2B).

iii Locomotor Activity and Open Field Test
Mice were placed in a dimly lit 40 x 40 cm open field (OF) chamber for 5 minutes to assess exploratory locomotor activity and anxiety-like behaviour. The activity was video recorded and Ethovision XT 8.5 analysis software (Noldus, Leesburg, VA) was used to assess the total distance traveled and time spent in the central or outer areas of the chamber.

iv Y-Maze
Working memory was evaluated using a two-trial Y-maze task. In the exposure trial, one arm of the Y-maze was blocked off so that the mouse was unable to enter ( Figure 2G). Mice were placed in one arm of the maze and allowed to explore for 12 minutes. One hour later, the block was removed, and mice were returned to the maze to explore all open arms for 5 minutes. The testing trial activity was video recorded and analysed using Ethovision 8.5 analysis software (Noldus, Leesburg, VA) to determine the distance traveled and time spent in each arm.

IV Metabolite Analysis
Thawed plasma samples were diluted with 5x 0.2% acetic acid and 1mM internal standards. Frozen brain tissue was diluted 3x with 0.2% acetic acid and 1mM internal standards and homogenized at 4° with 1.4mm zirconium ceramic oxide beads (Omni International) in an Omni International Bead Ruptor 24 Homogenizer (Kennesaw, GA; pulse duration: 45s, pulse number: 2, rest interval: 15s). The diluted brain homogenate and plasma were then transferred to Millipore Amicon Ultra Filters and centrifuged at 13500xg for 1 hour at 4°C. Liquid chromatography/mass spectrometry (LCMS) was performed on a Thermo Fisher Scientific Q Exactive mass spectrometer (Waltham, MA) with online separation by a Thermo Fisher Scientific Dionex Ultimate 3000 HPLC in the Mass Spectrometry Core Facility at the University of Texas Health San Antonio. Data were analysed using Xcalibur 2.2 software (Thermo Fisher Scientific).

V Statistical Analysis
All data were analysed using GraphPad Prism 8.2.0 (San Diego, CA). Data were considered significant if p < 0.05. Outlier data were identified using Chauvenet's outlier test and omitted from the analysis. Behavioural assays and metabolite concentrations were analysed by group using one-way ANOVA followed by Dunnett's post-hoc test. In the absence of significant main effects, post-hoc analysis was performed according to a priori hypothesis.

I Chronic Unpredictable Stress Drives Distinct Depressive-Like Behavioural Changes in C57Bl/6J Mice
Mice were tested in saccharin preference, nesting, open field test, and working memory Y-maze after 0, 10, 20, or 30 days of CUS exposure. CUS exposure failed to cause a reduction in saccharin preference in the acute 2-bottle choice task (Figure 2A). Nesting behaviour was increased in mice exposed to longer durations of CUS compared to low or no stress controls ( Figures 2B-2D). Mice exposed to 30 days CUS gathered nestlet pieces from more corners of the cage by the end of the assay and did so more rapidly. Distance moved and time spent in the central area of the open field test did not significantly differ between groups ( Figures 2E &  2F), suggesting that stress did not increase lethargy or drive an anxietylike phenotype. The Novel Arm Index of the working memory Y-maze task was significantly reduced after 10 days of CUS, however, indicating a reduction in working memory (p=0.015; Figure 2H). The Novel Arm Index remained reduced for 20 and 30 days of CUS but did not reach significance in these groups.

Figure 2:
Chronic unpredictable stress increases nesting behaviour and induces working memory deficits. Mice were assessed for depressive-like behavioural changes following 0, 10, 20, or 30 days of chronic unpredictable stress (CUS). A) Saccharin preference was measured for 2 hours following CUS and no effect was detected. B) 6 small nestlet pieces were arranged around the edges of the home cages of the mice to determine the effects of stress on nesting behaviour. Every 6 minutes, the nestlet pieces cleared from each corner of the cage were counted for a maximum score of 5 at each time point. C) Mice exposed to 30 days of CUS cleared significantly more corners at the end of the nesting assay than control mice. D) Nestlets were gathered significantly quicker by mice exposed to 20 and 30 days of CUS than 0 and 10 days of stress.

II Chronic Unpredictable Stress Increases Central Tryptophan Levels and Shifts the Peripheral Kynurenine Pathway Away from the Oxidative Branch
Brains and plasma were analysed by LCMS to determine if increasing exposure to CUS induces a shift in downstream tryptophan and kynurenine metabolism. Complete statistical analysis is reported in (Table 2). There was a significant main effect of CUS duration on central tryptophan, XT, and KYNA, and peripheral QA. Central QA was not reliably detected in any of the brain samples in this study. CUS caused a significant shift in the balance in downstream kynurenine metabolism. Complete statistical analysis of ratio data is outlined in (Table 3). In the brain, KYNA was elevated in relation to kynurenine after 10 and 20 days of CUS. Peripheral QA was significantly decreased over kynurenine and KYNA after 10 and 20 days of CUS exposure. These data suggest that CUS drives a shift towards KYNA and away from KMO-dependent metabolism.

III Chronic Stress Drives Serotonin Turnover in the Brain
There was a significant reduction of serotonin in the brain following 30 days of CUS compared to 0 days controls ( Figure 3). 5-HIAA was significantly increased at 10 days of CUS. The ratio of serotonin to tryptophan was significantly reduced for the duration of the study, while the ratio of 5-HIAA to serotonin remained elevated. These data indicate that CUS lowers central serotonin, possibly by up-regulating turnover to 5-HIAA.

Discussion
Chronic stress is a known risk factor for the development of depressive symptoms. The focus of this study was to characterize the change over time in downstream tryptophan metabolism pathways in conjunction with the depressive-like behavioural effects induced by stress. Here, we compiled several common stressors known to induce robust behavioural phenotypes or inflammatory processes into a chronic unpredictable stress (CUS) protocol [30][31][32][33][34]. We found that CUS significantly increased nesting behaviour and impaired working memory function ( Figure 3). These behavioural effects occurred in tandem with an elevation in central tryptophan, kynurenine, picolinic acid (PA), and 5hydroxyindoleacetic acid (5-HIAA). Central serotonin, xanthurenic acid (XT), and kynurenic acid (KYNA) levels were decreased. Peripherally, only quinolinic acid (QA) levels were reduced, while all others we measured remained steady.
The strain of mice used in this study (C57Bl/6J) is known to be resilient to the behavioural effects of stress compared to other strains [35][36][37][38][39]. In line with this, our behavioural assays did not reveal a robust depressivelike phenotype driven by our CUS protocol. In particular, we did not see the expected effects of CUS on anhedonia-like behaviour, a core component of the depressive-like phenotype, or on anxiety-like behaviour. This absence of a stress effect may be due to the insensitivity of the 2-hour saccharin preference test, which our lab typically performs over 24 hours. Further, while the open field test is widely recognized for its ability to detect anxiogenic effects by a reduction of the time spent in the central area and increase in thigmotaxis behaviour, studies have shown opposing effects of stress in this assay [40,41]. However, there was an impairment of working memory as early as 10-days of CUS, which remained impaired throughout stress exposure. We also saw a profound increase in nesting behaviour induced by stress. Nesting is an innate behaviour and serves as an indicator of well-being in mice [42]. The nesting clearance procedure was established by Negus et al. to measure pain-induced functional behavioural impairment; we hypothesized that chronic stress would likewise reduce nesting behaviour [29]. Instead, we saw that prolonged stress exposure increased nesting behaviour, as measured by the speed with which mice gathered the nestlet materials and the number of nestlets collected into one corner at the end of the assay (Figure 3). This increased activity, despite the absence of stress effects on locomotor activity, suggests that stress drives comfort-seeking behaviour. In this regard, nesting behaviour may serve as an alternative anxiety-like phenotype readout. It has been hypothesized that inflammation-induced depression is driven by a reduction of tryptophan in the brain due to increased kynurenine metabolism, thus limiting the tryptophan supply available for serotonin synthesis [43][44][45][46]. Stress similarly activates up-regulation of proinflammatory cytokines and drives imbalances in the kynurenine metabolism pathway [43,[47][48][49][50][51]. However, compared with inflammation, stress increases tryptophan concentrations in the brain [52][53][54][55]. This rise in tryptophan is accompanied by an up-regulation of serotonin turnover: serotonin is depleted while 5-HIAA concentrations rise, although conflicting results have also been reported with varying stressors and brain regions [48,[52][53][54][55][56][57]. Here we report similar trends in increased central tryptophan concentrations and serotonin turnover as a result of a modified CUS protocol (Table 2 & Figure 2). It has been proposed that stress drives the elevation of tryptophan in the brain by increased trafficking of tryptophan from the plasma in order to supply the source necessary for this increased serotonin turnover. Pre-treatment with valine, a competitor for the large amino acid transporter that allows tryptophan to cross the blood-brain barrier, prevents a restraint stressinduced elevation in tryptophan and downstream serotonin turnover in the brain [58]. Here, we show that peripheral tryptophan levels remain steady despite increased concentrations in the brain after stress ( Table  2).
Acute stress up-regulates the enzyme IDO, resulting in an increased tryptophan-kynurenine turnover [27,28]. We show here that while the concentration of kynurenine in the brain is increased after 10 days of CUS, the ratio of kynurenine to tryptophan did not reach significance. Unexpectedly, there was a main effect of stress-reducing central KYNA and XT and peripheral QA. The only downstream metabolite of the kynurenine pathway measured here that increased in the brain was PA. Interestingly, increased activity of ACMSD, the enzyme that converts 5hydroxyanthranilic acid (5-HAA) to PA, is implicated in depression in humans: increased production of PA is suggested to be protective against suicidality by reducing the availability of the precursor 5-HAA for QA metabolism [59]. The elevated kynurenine concentrations we measured here are shunted towards picolinic acid production, possibly driving the lack of depressive-like behaviours in these stress-exposed mice.