Attenuation of social interaction-associated ultrasonic vocalizations and spatial working memory performance in rats exposed to chronic unpredictable stress

Abstract

Exposure to unpredictable chronic mild stress (CUS) is a commonly used protocol in rats that is reported to evoke antidepressant-reversible behaviors such as loss of preference for sweetened water solution which is taken as an analog of the anhedonia seen in major depression. However, the induction of anhedonic-like behavior by chronic mild stress, gauged by an animal's preference for sucrose solution, is not fully reproducible and consistent across laboratories. In this study, we compared a widely used behavioral marker of anhedonia – the sucrose preference test, with another phenotypic marker of emotional valence, social interaction-associated ultrasonic vocalizations as well as a marker of an anxiety-like phenotype, novelty-suppressed feeding, and cognitive performance in the eight arm radial maze task in adult male Sprague-Dawley rats. Chronic four-week exposure to unpredictable mild stressors resulted in 1) attenuation of social interaction-associated ultrasonic vocalizations 2) attenuation of spatial memory performance on the radial arm maze 3) attenuation of body weight gain and 4) increased latency to feed in a novelty-suppressed feeding task. However, chronic exposure to CUS did not result in any significant change in sucrose preference at one-week and three-week intervals. Our results argue for the utility of ultrasonic vocalizations in a social interaction context as a comparable alternative or adjunct to the sucrose preference test in determining the efficacy of CUS to generate an anhedonic-like phenotypic state.

1. Introduction

It has been well established that chronic emotional stress plays a pivotal role in the genesis of many psychiatric disorders with induction of both short-lasting and long-lasting alterations in behavior and physiological functions [1, 2]. Such emotional stressors are one of the main sources of stress in human life, especially for those low in the social hierarchy, and play a major role in the pathogenesis of anxiety and depressive disorders [3, 4]. In social settings, stress can occur throughout the lifespan, and can range from childhood neglect to peer abuse such as school bullying in adolescence or workplace harassment in adulthood [5, 6]. Furthermore, chronic stress may be associated with fearful (and life-threatening) events of traumatic nature such as violence, war, injury or assault [7].

There are very few animal models of human major depressive disorder (MDD) that can adequately provide the face, construct and predictive validity for bench research to be translated into bedside application. Mild Chronic Unpredictable Stress (CUS) is one such paradigm that has been used to model the human symptom of anhedonia (defined as loss of interest in daily activities that were previously enjoyable) [8]. While a sufficient volume of literature supports the efficacy of CUS in altering observable behaviors, sucrose testing is by far the most cited measure of ‘anhedonia’ [9]. The use of sucrose testing (of either consumption or preference) generalizes palatable taste reactivity as an index of hedonic state in laboratory rodents. However, reduction in preference for a sweetened water solution has been widely criticized as being unreliable due to variable responsiveness of rodents to CUS [10] and having possible relationships to body weight and nutritional status as well as caloric intake which may confound the results [11]. In fact, these observed effects may be limited to the gustatory circuitry and metabolic demand.

A large body of evidence has accumulated over the last two decades that highlights the difficulties associated with sucrose testing as a consistent measure of hedonic drive [12]. For example, chronic stress has no effect on sucrose consumption under a progressive ratio schedule as would be expected if hedonic drive were reduced [13]. Similarly, sucrose consumption was more dependent on food deprivation and weight changes seen with chronic stress paradigm, and independent of other elements of stress protocol [14]. While CUS reproduces characteristic behavioral responses, there is a need to validate objective alternatives to sucrose testing as a measure of a rodent's affective state. The goal of the present study was to compare the effects of CUS on sucrose preference to other behavioral markers relevant to major depressive disorder in order to increase the reproducibility of the paradigm.

Therefore, in Experiment A, we examined multiple measures to seek more reliable behavioral markers for the effects of CUS, starting with the sucrose preference test. Based on previously published studies, it was expected that the stressed rats would show a reduced preference for the sucrose solution [15]. However, given that sucrose preference has also been shown to paradoxically increase or remain unchanged with exposure to chronic stress, we also tested other behavioral markers of chronic stress in addition to sucrose preference test [16]. These markers included tests of body weight gain, separation/anticipatory ultrasonic vocalization and novelty-suppressed feeding. A total of 20 rats were used for this experiment (10 per treatment group).

Experiment A was followed by Experiment B with an additional test of a behavioral marker relevant to MDD, namely, spatial working memory in an 8-arm radial maze [17]. Spatial memory is a type of hippocampal-dependent learning and memory, and evidence suggests frequent impairment of hippocampal-dependent learning and memory in MDD [18]. A total of 30 rats were used for this experiment (14 control and 16 CUS). In total, 50 rats were used in Experiments A and B.

2. Materials and Methods

2.1. Materials and Methods – Experiment A

The goal of the first experiment was to identify reliable behavioral markers for the effects of mild chronic unpredictable stress (CUS). The markers included assessment of gain in body weight, sucrose preference, novelty-suppressed feeding, and anticipatory ultrasonic vocalization after brief separation from cage-mate.

2.1.1. Animals and Housing

For both Experiments A and B, adult male Sprague-Dawley rats (Charles River, Wilmington, Mass., USA) weighing 200-250 g at the start of experiment were housed two per cage (25×48×20 cm) in a temperature and humidity-controlled colony room (~21°C, 40–50%) at the University of Mississippi Medical Center Laboratory Animal Facility. The rats were maintained on a reverse 12:12 light/dark cycle, with lights off at 0700 h. All behavioral testing took place during the dark phase of the cycle. Food and water were available ad libitum, except during testing. Before the beginning of experiments, all animals were handled for approximately 15 min, daily for 3 days. All procedures were approved by the University of Mississippi Medical Center Institutional Animal Care and Use Committee and conformed to the guidelines of the National Institutes of Health.

2.1.2. Sucrose Preference

Prior to beginning of testing, rats were habituated to the presence of the two drinking bottles for 5 days (4 hours each day) in their home cages. One of the bottles contained sucrose in increasing concentrations each day (0.1, 0.3, 1.0, 3.0, 10%). This allowed us to determine that a 3% sucrose solution consistently (S.E.M < 10% of mean) elicited a 3-fold preference over tap water while the 1% solution did not reliably maintain a strong preference and the 10% solution elicited a larger, but much more variable preference (S.E.M. ~20% of the mean). Therefore, for all subsequent studies, we compared the preference for a 3% sucrose solution to tap water. Following this acclimation, rats had the free choice of either drinking the 3% sucrose solution or tap water for a period of 3 consecutive days (4 hours each day). Sucrose preference was calculated as a percentage of the volume of sucrose intake over the total volume of fluid intake [11] and analyzed over the testing period of 3 days via two-factor ANOVA followed by Tukey's HSD for between group comparisons. There was a significant difference in sucrose preference between the three days [F_(2,38)=4.316, p<0.05]. Post hoc comparisons using the Tukey HSD test indicated that the mean (M) % preference score for the Day-1 sucrose preference (M=80.63, SD= 11.59) was significantly different from the Day-2 sucrose preference (Mean=71.99, SD= 7.84). However, the Day-3 sucrose preference (Mean=76.74, SD= 11.64) did not significantly differ from either Day-1 or Day-2 sucrose preference. Since the sucrose preference differed significantly between the first and second day only, a two-day protocol of sucrose preference testing was employed during the CUS treatment.

2.1.3. Mild Chronic Unpredictable Stress (CUS)

Rats were assigned to one of two groups (n=10 each per group) and were either exposed to mild chronic unpredictable stress (CUS) or handled to serve as no stress controls. Rats in each group were matched as closely as possible for body weight, baseline sucrose preference and total fluid intake. Rats assigned to the CUS group were exposed to the CUS protocol shown in Table-1. This 10-day protocol was systematically repeated to maintain the element of unpredictability throughout the experiment and for a total of 35 days (5 weeks of CUS treatment). During this period, control animals were regularly handled, weighed and housed separately without any exposure to the CUS paradigm.

TABLE 1.

Riaz et al., Attenuation of social interaction-associated ultrasonic vocalizations and spatial working memory performance in rats exposed to chronic unpredictable stress.

Day	Stressor 1	Stressor 2
Day 1	50 min cold room	60 min cage rotation
Day 2	4 h wet bedding	12 h lights on during dark cycle
Day 3	60 min restraint stress	3 h lights off during light cycle
Day 4	50 min cage rotation	15 h food and water deprivation during the dark cycle
Day 5	15 min cold room isolation	17 h isolation housing during light cycle in clean cage
Day 6	4 h wet bedding	3 h lights on during dark cycle
Day 7	30 min cage rotation	1 h lights on during dark cycle
Day 8	5 min swimming exposure during dark cycle	60 min restraint stress
Day 9	4 h wet bedding	12 h food deprivation during dark cycle
Day 10	45 min cold room isolation	6 h lights on during dark cycle

CUS treatment paradigm – 2 stressors/day repeated daily: This 10-day protocol was systematically repeated to maintain the element of unpredictability throughout the experiment and for a total of 35 days (5 weeks of CUS treatment) for Experiment-A, and 28 days (4 weeks of CUS treatment) for Experiment-B. During both experiments, control animals were regularly handled, weighed and housed separately without any exposure to the CUS paradigm.

Body weight was measured before CUS (Day 0), 2 days after the beginning of CUS (Day 2) and every 4 days thereafter. Sucrose preference testing was conducted after 1 and 3 weeks exposure to CUS. USVs were tested 1 to 5 days after the conclusion of CUS and novelty-suppressed feeding 24 hours after the last USV recording.

2.1.4. Sucrose Preference Testing

Sucrose preference testing was carried out in a quiet dimly lit testing area. The animals were transported from the colony, separated into two clean cages and allowed to acclimate for 2 h during week-1 and week-3 of CUS treatment. After two hours acclimation, rats were presented with a choice of 2 bottles during the 2 consecutive test days (4 hours each day) in both week-1 and week-3. One bottle contained tap water, and the second contained a 3% sucrose solution. Water and sucrose solution intake was measured daily, and the position of the two bottles was randomly switched to reduce any confound produced by a side bias. Sucrose preference was calculated as outlined above and analyzed over the testing period of 2 days via single factor (CUS exposure), repeated measures ANOVA followed by planned comparisons using uncorrected univariate F-tests for between-cell comparisons.

2.1.5. Body Weight Analysis

During the course of CUS treatment (35 days), animals from both the control group and CUS-treated group were weighed every third day (at 09:00 h) to monitor their overall health. The difference in weight gain between the two groups was analyzed by single factor (CUS exposure) repeated measures ANOVA, and followed by planned comparisons using uncorrected univariate F-tests for between-cell comparisons.

2.1.6. Anticipatory USV Response Test

2.1.6.1. USV Recording Apparatus

The USV recording apparatus consisted of four Plexiglas experimental chambers constructed of 2.5 cm thick black Plexiglas. After preliminary studies identified a sound echo, a layer of sound absorbing foam approximately 5 mm thick was applied to the chamber walls which reduced echo to low levels that permitted manual evaluation of the sonogram. Each chamber was equipped with a ventilation fan which also emitted < 20 dB white noise with no detectable ultrasonic noise. In addition, each chamber was equipped with a 6 W fluorescent light which was masked to produce red light. High-sensitivity ultrasonic microphones (Sonotrack, Metris B.V., KA Hoofddorp, Netherlands) with a frequency response range of 15 to 125 kHz were securely placed via small sealed holes in the top panels of each chamber. The main Sonotrack unit was connected to a Dell Computer workstation running the Sonotrack software in a Microsoft Windows environment. The entire apparatus was placed in a dedicated sound-attenuated room. Ultrasonic vocalizations were measured using Sonotrack ultrasonic microphones, bandpass filtered and digitized by Sonotrack A/D conversion card and stored as .dat-files (internal Sonotrack format for sound data storage) and .wav-files (external storage format for media player usage) for later analysis. Sound analysis was performed using Metris Sonotrack software (Sonotrack, Metris B.V., KA Hoofddorp, Netherlands). Although Sonotrack is equipped for automatic detection of USVs, we found that significant numbers of USVs were missed in a background noise of undetermined origin and therefore decided to complete all analyses manually. Sound files were converted to spectrograms for manual visual counts of calls. A sound was considered to be a 50 kHz USV if it fell between 35 and 70 kHz with duration of 30-80 ms. Calls were considered in the 22 kHz category if they fell between 18 and 32 kHz with duration of 300-3000 ms. All sound files were manually scored by trained experimenters without knowledge of test group assignment.

2.1.6.2. Anticipatory USV Response Testing

Anticipatory Ultrasonic Vocalization (USV) response testing was carried out for 4 days following the 28-day CUS exposure paradigm, and involved both the control and the stressed animal groups. On the day prior to test onset, CUS-treated and control animals were placed individually in the test chamber for 30 min to habituate them to the testing situation. On test days, experimental subjects (CUS-treated or control) and cage-mate partners were individually isolated in holding cages for 4 h prior to testing to assess social USVs. On each of the 4 test days, cage-mates were taken from their holding cages and each animal was placed alone into one of the recording chambers for a 5-min period. One control pair and one CUS-treated pair were tested at a time and the assignment of testing chambers was randomized. After 5 minutes of recording individual animals alone, the cage-mate partners were reunited in one of their test chambers, and the dyad was allowed to interact for 5 minutes. USVs were recorded during the anticipatory (5 minute) period on test days 1,2,3, and 5 and the (5 minute) social interaction period on test days 1 and 5 for this experiment and the digitized recordings were saved for later analysis. Day 4 of testing is omitted due to construction near the testing area which precluded recording. Unless otherwise stated, all animals were returned to their home-cage immediately after testing on each of the 4 test days. After examining automatic counting in low, middle and high frequency bands, it became clear that rats were emitting a significant number of calls of low intensity that were not picked up from the background noise and therefore were not automatically counted. We therefore decided to manually count all sound files. Initially, individual calls were counted above and below 45 kHz to represent high frequency (~55 kHz) and low frequency (~22 kHz) vocalizations. However, close examination of the resulted data set indicated that few if any vocalizations were emitted below 40 kHz and all were of short duration (<100 msec) indicating that they did not have the characteristics of the low frequency vocalizations often attributed to distress. Therefore all vocalizations were counted as high frequency vocalizations. After all sound files had been manually scored as described above, a single factor ANOVA was used for statistical analysis followed by planned comparisons using uncorrected univariate F-tests for between cell comparisons. The number of daily vocalizations, particularly among the control groups were fairly variable and so were transformed as log (raw value +1). Aggregate call data over all four days was analyzed with a single factor (CUS) repeated measures ANOVA (trials) followed by comparison between the two groups using a two-tailed Fisher's Exact test.

2.1.7. Novelty Suppressed Feeding

The Novelty Suppressed Feeding (NSF) test was carried out at the end of last and final week 5 of this experiment, following the 4 days of anticipatory USV response testing. Prior to the NSF test, the rats were food deprived overnight (~16 hours). On the day of testing, 4 hours after lights off, the animals were presented with a single food pellet (chow) placed in the middle of a novel environment, which was a circular enclosure of 100cm diameter with clean bedding. Each of the rats was placed near the enclosure wall and allowed to explore freely for 10 minutes, and the latency to start feeding (during the 10 minute assay) was recorded. The enclosure was thoroughly wiped down between subjects with a 10% ethanol/water solution to eliminate scents and olfactory traces. The difference in latency to feed between the two groups was analyzed via Student's t-test, with statistical significance set at α of 0.05.

2.2. Materials and Methods – Experiment B

The goal of the second experiment was to establish the effects of CUS on hippocampal-dependent spatial memory performance.

2.2.1. Spatial (hippocampal-dependent) Memory Performance on Radial Arm Maze

2.2.1.1. Radial Arm Maze Apparatus

The Plexiglas radial maze was elevated 30 cm off the ground, and consisted of 8 arms each measuring 10×80 cm with black flooring and clear walls. Arms were connected by a round central platform 35 cm in diameter. Each arm contained a depressed circular food cup at its terminal end, which was baited during testing with one 45 mg bacon-flavored sugar pellet (Purina TestDiet #1813244, Richmond, IN). The principle behind testing on a radial arm maze apparatus is that an animal quickly learns that there is only one pellet in each arm and thus avoids the already visited and de-baited arm. The animal keeps on exploring all the unexplored arms until all of the arms have been visited. The animal uses spatial cues that surround the maze to remember which arms it has entered, and hence radial arm maze testing is used to test for intact spatial memory.

2.2.1.2. Spatial Working Memory Training

Prior to CUS treatment, animals were trained on the radial arm maze task for 3 weeks with one daily trial during the dark cycle. At the beginning of each trial, the rat was placed in the central platform and was allowed to freely move on the maze until either all eight arms were entered at least once, or until 300 seconds had passed; which ever came first. An arm entry was recorded when all four of the animal's legs had crossed the demarcated threshold of the arm.

Performance was scored for the number of correct (unvisited) arm entries in the first eight arm visits. Success rate for both training and testing trials was calculated as the percentage of correct entries in first eight visits (Previously unvisited arms entered/first eight arm visits × 100). All the animals achieved the acquisition criteria of 80% success rate by the 18^th trial (1 trial/day) as shown in Figure-6A.

Figure 6.

A) Radial Arm Maze training: All the animals achieved the acquisition criteria of 80% success rate by 18^th trial (1 trial/day). Open diamonds connected by a dotted line represent the mean of all animals before dividing them into the control and CUS groups. Closed squares and closed diamonds connected by solid lines represent animals divided into control (CTRL) and chronic unpredictable stress (CUS) groups. B) Radial Arm Maze testing: CUS-treated animals had a significant reduction in memory retrieval during all 4 weeks of CUS treatment, when compared with pre-treatment performance in the radial arm maze testing [F_{(4, 40),} P < 0.05, *].

2.2.1.3. CUS treatment and Spatial Working Memory Testing

Following the training period, animals were assigned to either CUS-treatment (n=16) or no stress control (n=14) groups. Rats in each group were matched as closely as possible for body weight. Animals from both CUS-treated and control groups were tested for spatial working memory performance on the radial arm maze, once every week during the 4 weeks of CUS treatment. The spatial memory performance over of the course of testing period was analyzed via single factor (CUS) ANOVA with repeated measures (trials), followed by planned comparisons using uncorrected univariate F-tests for between cell comparisons.

2.3. Statistical Analysis

All data were analyzed as described in 2.1.2., 2.1.4. – 2.1.7, and 2.2.1.3. using IBM SPSS version 22. Significance for all tests was defined as α or p value < 0.05.

3. Results

3.1. Experiment A

3.1.1. Body Weight

The control animals showed significant weight gain, compared to baseline (Day 0) from day 15 to 33, whereas no such weight gain was observed in CUS-exposed animals [F_(9,162)=12.95, p <0.05, Figure-1]. Between-group analyses showed that CUS-exposed animals weighed significantly less compared to control animals through Day-26 till Day-33 (Week-5) of CUS treatment. In addition, weight of CUS animals at Day-33 was similar to before CUS, suggesting that they did not gain weight. In contrast, the control animals showed a significant gain in weight at Day-33 compared to before CUS [t₍₁₉₎=2.09, p <0.05, Figure-1].

Figure 1. Weight Change.

Posthoc analysis showed that control animals gained significant weight, compared to CUS-treated group during the CUS-treatment period (Day-10 to Day-33), [F_(9,162)=12.95, p <0.05, *]. Control animals also showed a significant weight gain on Day-33 compared to pre-CUS Day-0 [t₍₁₉₎=2.09, p <0.05, §].

3.1.2. Sucrose Preference Test

CUS-treated rats did not significantly differ from control group in sucrose preference at the end of week-1 of CUS treatment, or week-3 of CUS treatment [F_(1,38)=0.02, n.s., Figure-2A-B].

Figure 2. A) Week-1 of CUS.

CUS-treated rats did not differ from control group in sucrose preference at the end of week-1 of CUS treatment [F_(1,38)=0.02, n.s.] B) Week-3 of CUS: CUS-treated rats did not differ from control group in sucrose preference at the end of week-3 of CUS treatment either.

3.1.3. Novelty Suppressed Feeding Test

CUS-treated rats showed a significant increase in latency to feed during the testing period compared to the control group [10 min; CUS vs. control, [t₍₁₉₎=2.09, p<0.05, Figure-3].

Figure 3. Novelty suppressed feeding (NSF) test.

CUS-treated rats showed a significantly increased latency to feed during the testing period [10 min; CUS vs. control (CTRL), t₍₁₉₎=2.09, P<0.05, *].

3.1.4. Anticipatory Ultrasonic Vocalizations

CUS-treated rats produced fewer total anticipatory USVs than control rats through all 4 days of testing, with a trend nearing significance for between-subjects group effect [F_(1,18) = 4.115, p = 0.058, Figure-4A]. Pair-wise comparisons between the two groups across all 4 testing days showed a significant reduction in total anticipatory USVs in CUS-treated rats vs. control rats on Day-1 of testing [F_(1,18) = 7.507, p < 0.05, Figure-4A], and a trend towards reduction in total anticipatory USVs on Day-3 of testing, however the observed difference between the 2 groups did not achieve statistical significance [F_(1,18) = 3.595, p = 0.08, Figure-4A]. During the 4 testing days, anticipatory USVs were positively recorded for 28 out of 40 testing sessions for control animals, while CUS-treated animals emitted detectable anticipatory USVs in only 15 out of 40 testing sessions [Figure-4B]. Comparing the 2 groups for only the testing sessions with detectable emitted USVs using a two-tailed Fischer's Exact test revealed a significant difference in USV production between the two groups p < 0.05, Figure-4B. There were no Main or Interaction effects on average call duration and even though CUS animals emitted consistently shorter USVs than controls, this did not achieve statistical significance [F_(1,18) = 3.474, p = 0.08]. When cage mates were reunited for 5 min on the first or fifth testing day, CUS pairs consistently emitted less than half the number of calls of unstressed controls [One tailed t₍₈₎ = 2.128, p = 0.033] Figure 5.

Figure 4.

A) 50-kHz USVs: CUS-treated rats produced significantly fewer anticipatory 50 kHz USVs (Band-3; 44.5-100 kHz) through all 4 days of testing [F_(1,8) = 7.20, p < 0.05,*]. B) Total USVs: CUS-treated rats emitted significantly fewer total USVs, with all bands combined [F_(1,8) = 7.20, p < 0.05,*].

Figure 5.

50-kHz USVs of Reunited Cage Mates. CUS-exposed rats produced significantly fewer 50 kHz USVs than controls (CTRL) (Band 3: 44.5-100 kHz) when reunited with cage mates on days 1 and 5 of testing [One tailed t(8) = 2.128, p = 0.033].

3.2. Experiment B

3.2.1. Effect of CUS on memory retrieval

CUS-treated rats showed progressively more errors every week in radial arm maze testing during the 4 weeks of CUS treatment compared to the pretreatment baseline. In contrast, the performance of the control animals was unchanged over those four trials when compared to baseline. Although the error rate of CUS-treated rats and control rats did not differ at baseline, by the second week of CUS treatment, there was a significant increase in errors in CUS-treated rats compared to control rats that persisted throughout the rest of testing. Single factor (Treatment group) repeated measures (trials) ANOVA revealed a significant main effect of trials [F_(4,112) = 9.789, p < 0.001] and of group [F_(1,28) = 9.570, p < 0.001] as well as a treatment group by trials interaction [F_(4,112) = 9.732, p < 0.001, Figure-5B]

4. Discussion

This study used the stress exposure paradigm of CUS, and compared multiple behavioral measures for their predictive efficacy as markers of chronic stress. The CUS-treated animals showed attenuation of social interaction associated USVs, attenuated weight gain, suppressed feeding when exposed to novelty, and attenuation of spatial memory. The stress-induced reduction in weight gain is in accordance with previously published studies, which show a similar effect [19, 20]. One possible explanation for this decrease in weight gain could be the use of food deprivation as one of the unpredictable mild stressors. However, not only was food deprivation stress an infrequent event during CUS, evidence suggests that decreased hedonic responsiveness following CUS is not secondary to loss of body weight [21].

In the first experiment of this study, there was no reduction in sucrose preference between CUS-treated and control animals. Furthermore, the duration of stress exposure did not change sucrose preference as CUS-treated animals showed no change in preference at end either the first week or third week of treatment. Our observation supports previous studies that have reported difficulty in eliciting a decreased sucrose preference using multiple variants of chronic stress procedures [11, 14, 22-24]. These inconsistencies have been recorded in chronic stress models of both mice and rats, and such equivocal and even contradictory findings weaken the reliability of sucrose preference testing as a behavioral marker of hedonic state [1, 25-27]. Interestingly, however, recent studies by Wiborg and coworkers [10] suggest that these inconsistencies may result from a differential population change in sucrose preference during CUS exposure with some animals displaying profound reductions, while other display either no or very little response to stress. If this is replicable, it would lend population validity [28, 29] to CUS as a model of depression since not all individuals subjected to stress develop major depressive disorders.

Sucrose preference testing was used for this study, since previous studies have shown that sucrose consumption is a weak and inconsistent index of reward responsiveness in chronic stress paradigm [11]. In fact, previous studies show that any observed effect of reduced sucrose consumption in CUS-exposed animals is lost, when corrected for stress-induced weight changes [14]. Thus, the sucrose preference test was chosen here rather than the sucrose consumption test as a behavioral marker of CUS-induced anhedonia. Sucrose preference was calculated using the previously published method [11, 30-32]. Since no change was observed in hedonic responsiveness of sucrose preference; it is likely that reduction in the consumption of palatable solutions alone is insufficient to be equated with anhedonia. We therefore tested for hedonic drive through ultrasonic vocalization (USV) analysis in a social behavior paradigm, as sucrose palatability only represents one limited facet of an animal's hedonic state.

Rodent USVs have been well-characterized for objectively representing the affective state of a rodent (reviewed in [33]). In fact, the strong correlation between measurable indices of autonomic nervous system (anxiogenic behavior) and USVs was well-characterized before wide scale adaption of sucrose preference analysis as an indicator of affective state [34]. Unlike strain-specific differences in sucrose preference that vary by genetic and metabolic features, all rodent vocalizations can be objectively characterized by acoustic parameters and frequency modulation [35, 36]. These broadly comparable features of rodent communication system (USVs) can be objectively analyzed across laboratory rat and mice strains to separate unique acoustic features in context of unique behavioral paradigms [37-39]. Furthermore, detailed qualitative analysis of adult rodent vocalizations in social context is being explored for behavioral phenotyping of rodent models of human neurodevelopmental and neuropsychiatric disorders [40].

In comparison with sucrose preference test, USVs essentially serve as a ‘self-reporting feature’ of a rodent's affective state [41], and adult male Sprague-Dawley rats emit characteristic anticipatory USVs in response to social stimuli [42, 43]. Anticipatory USVs were assessed in a cage-mate separation paradigm to determine the efficacy of CUS in generating depressive-like behaviors. In this study, the stress-exposed dyads emitted very few 50 kHz vocalizations as opposed to their control counterparts. Previous studies have shown a strong association of 50 kHz USVs with pleasurable and rewarding activities [33, 44, 45]. The decrease in anticipatory 50 kHz vocalizations observed during the 5^th week of CUS exposure, with no significant variation across the testing days, provides a comparable index of an animal's anhedonic-like affective state.

A potential limitation of using any measure of reward-seeking behavior, including sucrose preference and possibly, USV emission, as an index of anhedonia is that it can be difficult to separate the motivational aspect of reward seeking from the animal's perception of the value of the reward. Richardson and Roberts have argued that progressive ratio schedules of reinforcement are more influenced by motivation than by reward perception [46]. In that light, the work of Barr and coworkers demonstrating that CUS has no effect on progressive ratio responding for sucrose [13] would argue that CUS affects reward perception more than motivational salience. Nonetheless, since reward perception influences motivation, this data should not be considered as conclusive evidence of the induction of anhedonia by CUS and further studies are warranted to tease these issues apart.

Major depressive disorder is often characterized by cognitive deficits (reviewed in [47]) which can be dissociated from other features of MDD such as fatigue, withdrawal or anhedonia. Because deficits in cognition are such a common element of MDD, several groups, including ours, included a test of learning and memory in evaluating the effects of CUS [19, 48-50]. While multiple laboratories have demonstrated that CUS impairs acquisition of a reference learning task either when the animals are subjected to CUS shortly before or during the learning task, only one group [51, 52] has presented evidence suggesting that exposure to CUS impairs performance in a working memory task (spontaneous alternation). However, even their work was limited to a single test during the CUS exposure.

In contrast to the work of Henningsen and coworkers [51, 52], our studies repeatedly exposed rats to the eight arm radial maze during chronic unpredictable stress. Thus, a particularly significant finding of the present work is that CUS clearly produced a progressive deficit in working memory performance in the eight arm radial maze. Moreover, our data demonstrates that working memory is impaired after even one week of CUS and appears to become progressively worse with continued exposure to CUS.

Our work in tandem with that of Henningsen and colleagues [51, 52] clearly demonstrates that CUS produces deficits in both working memory and acquisition of reference memory information. It is of interest to note that none of the previous studies demonstrating that CUS interferes with reference memory acquisition began learning and memory testing before at least two weeks of CUS exposure. Our data on the progressive nature of CUS-induced impairment of working memory suggests that these working memory deficits may cause or contribute to the later impairment of acquisition of reference memory tasks. It will be important to examine, in future studies, whether working memory deficits persist after CUS exposure is terminated and for how long and also if and how quickly these deficits respond to antidepressant treatments.

In this manuscript, we have presented novel evidence that the CUS model of MDD-like symptoms includes both deficits in cognitive function (working memory) and in response (55 kHz USVs) to situations that normally possess a positive emotional valence such as anticipation of reuniting with a cage mate or actual interaction between cage mates. While clearly more work is needed to determine the time-course and antidepressant-sensitivity of CUS on these measures, they offer a robust and, in the case of CUS effects on working memory, early indicator of the behavioral effects of this paradigm. Moreover, the basic neural circuity underlying both chronic stress and working memory [53] and the emission of 55 kHz USVs [33] has been elucidated which may suggest novel diagnostic therapeutic targets for major depressive disorder.

Highlights.

• CUS exposure did not alter sucrose preference in adult male Sprague Dawley rats

• CUS exposure attenuated social interaction-associated ultrasonic vocalizations

• CUS exposure attenuated cognitive performance on a spatial working memory task

• CUS exposure resulted in attenuation of weight gain in CUS-treated rats

• CUS exposure resulted in increased latency to feed in a novel environment