Early life stress induces hyperactivity but not increased anxiety-like behavior or ethanol drinking in outbred heterogeneous stock rats

Abstract

Early life stress is known to impact vulnerability to psychopathological disorders in adulthood, including anxiety and alcohol use disorder (AUD), but the mechanisms underlying susceptibility to these outcomes are not fully understood. In the current study, we used adolescent social isolation (ASI) to determine whether Heterogeneous Stock (HS) rats, an outbred model used for genetic fine-mapping, could be used to study the genetics contributing to ASI-induced anxiety- and AUD-like behavior. We isolated (ASI) or group-housed (adolescent group-housed; AGH) 64 male HS rats at 4 weeks of age. After 5 weeks in these housing conditions, multiple anxiety and coping/despair-like behaviors were measured. All rats were then individually housed and assessed for voluntary ethanol self-administration. At euthanasia, synaptoneurosomes were isolated from a subset of brains to examine the expression of two proteins associated with alcohol drinking-related behaviors, GluA1 and SK2, in the dorsal (dHC) and ventral hippocampus (vHC). We found that ASI increased hyperactivity in the open field test relative to AGH, with no changes in other anxiety-like behaviors. Surprisingly, ASI rats demonstrated decreased immobility and increased climbing in the forced swim test relative to AGH. In contrast to prior studies by us and others, we found no difference in self-administration of 20% ethanol, with decreased ethanol self-administration in ASI relative to AGH rats at higher ethanol concentrations. Furthermore, while ASI in Long-Evans rats resulted in decreased SK2 expression in vHC synaptosomes, no differences were seen in vHC synaptosomes for SK2 or GluA1 in HS rats. These results demonstrate that HS rats are protected against many of the negative effects previously seen in response to ASI, namely anxiety-like behavior and increased ethanol self-administration. The current work suggests that a lack of change in SK2 and GluA1 expression levels in the vHC may play a role in conferring this protection.

Introduction

Early life stress (ELS) during adolescence is known to have lasting effects impacting vulnerability to psychopathological disorders in adulthood, including post-traumatic stress disorder (PTSD) (Peleikis, Mykletun, & Dahl, 2004), anxiety (Fonzo et al., 2016), and alcohol use disorder (AUD) (Muller et al., 2015), among other outcomes. The mechanisms by which ELS affects an individual’s susceptibility to these psychopathological outcomes are not fully understood, and rodent models have been used to address this. Adolescent social isolation (ASI) in rodents has been shown to induce enduring increases in anxiety-like behaviors and ethanol self-administration (Chappell, Carter, McCool, & Weiner, 2013; Lopez, Doremus-Fitzwater, & Becker, 2011; Yorgason, Espana, Konstantopoulos, Weiner, & Jones, 2013). Studies focused on the mesolimbic circuitry have identified areas in which neurobiological adaptations may be induced by ASI (Butler, Karkhanis, Jones, & Weiner, 2016), including the ventral tegmental area (Fabricius et al., 2010; Koob & Volkow, 2010; Whitaker, Degoulet, & Morikawa, 2013), the nucleus accumbens (Karkhanis, Rose, Weiner, & Jones, 2016; Yorgason et al., 2013, 2016), and the basolateral amygdala (Karkhanis, Alexander, McCool, Weiner, & Jones, 2015; Rau, Chappell, Butler, Ariwodola, & Weiner, 2015). Most recently, we have generated convergent evidence in Long-Evans rats that ASI and a rodent model of alcohol dependence both lead to a decrease in hippocampal glutamatergic synaptic transmission that is largely restricted to the ventral hippocampus (Almonte et al., 2017; Ewin et al., 2019). Notably, the ventral hippocampus is known to play an integral role in regulating anxiety-like behaviors (Strange, Witter, Lein, & Moser, 2014).

There is also evidence suggesting that genetics may play a role in susceptibility and resistance to ELS-induced disorders, although few specific genes have been identified (Lee, Oswald, & Wand, 2018). Using ASI in the Long-Evans outbred rat, our group has found a high degree of variability in both anxiety-like behaviors and ethanol self-administration (Chappell et al., 2013). One potential explanation for this variability is genetic differences in susceptibility to the behavioral and neurobiological effects of ASI. Here, we sought to utilize outbred Heterogeneous Stock (HS) rats, a model often used for genetic fine-mapping of complex traits, to test this hypothesis.

The HS rat population was created through the outbreeding of eight inbred rat strains (Hansen & Spuhler, 1984). After 80 generations of breeding, the genetic diversity is such that genetic fine-mapping is achievable to regions of a few megabase pairs (Solberg Woods & Palmer, 2019). Our laboratory and others have previously used the HS population to fine-map multiple traits such as diabetes, obesity, and fear-related anxiety behaviors (Chitre et al., 2020; Rat Genome Sequencing and Mapping Consortium, 2013; Keele et al., 2018; Tsaih et al., 2014; Solberg Woods et al., 2012; Solberg Woods, Holl, Tschannen, & Valdar, 2010). The current study was conducted to determine whether HS rats would be a useful model to identify genetics that underlie susceptibility to ASI-induced increases in anxiety-like behavior and ethanol self-administration. We hypothesized that ASI would increase anxiety-like behavior, induce hyperactivity in response to a novel environment, and increase ethanol self-administration in HS rats.

Methods

Heterogeneous stock rats

The HS rat colony was established in 1984 by the NIH by breeding together eight inbred founder strains: ACI/N, BN/SsN, BUF/N, F344/N, M520/N, MR/N, WKY/N, and WN/N, and maintaining the colony in a way that minimizes inbreeding (Hansen & Spuhler, 1984). The HS colony was maintained at the Medical College of Wisconsin beginning in 2006, with a colony set up at Wake Forest School of Medicine (WFSM) in 2017. The animals in the current study come from the WFSM colony that has been named NMcwiWFsm:HS (Rat Genome Database number 13673907). At the time of this study, the colony was maintained using 64 breeder pairs in a way that takes into account family relationships using a kinship coefficient to ensure that only distantly related animals are mated. Sixty-four male HS rats were used in the current study. Housing rooms were maintained on a 12-h light/dark cycle (7:00 AM to 7:00 PM). Housing conditions (e.g., social isolation or group housing) are described below. To maximize genetic diversity, we used up to four males per breeder pair and assigned no more than two males per housing condition from each breeder pair. All protocols were approved by the WFSM IACUC committee.

Adolescent social isolation

The methods we used mimic previously established protocols for ASI in Long-Evans rats followed by testing for anxiety-like behaviors and intermittent home-cage ethanol self-administration (Chappell et al., 2013). Male HS rats were weaned at 3 weeks of age and placed into groups of four rats per cage (612 mm × 435 mm × 216 mm; 2065 cm2). Prior to being placed in a housing condition, we ran a forced swim test, as described below, at 3.5 weeks of age to determine whether adolescent immobility correlated with future response to ASI. At 4 weeks of age, half of the rats were isolated (ASI, n = 32) into individual cages (259 mm × 476 mm × 209 mm; 910 cm2) and half remained in the group housing condition (AGH, n = 32). Rats remained isolated or group housed for 5 weeks, after which time behavioral testing was performed to assess anxiety-like (elevated plus maze, elevated zero maze, open field test) and despair-like (forced swim test and splash test) behaviors. Rats remained in the assigned housing condition throughout the emotional behavioral testing phase. Tests were performed 1 week apart, and the order was chosen to minimize stress on the animal to avoid influences of test order on behavioral outcomes (McIlwain, Merriweather, Yuva-Paylor, & Paylor, 2001). After all emotional behavioral testing, rats were placed into individual cages and ethanol self-administration was administered as described below. Rats were run in two groups of 32 (16 ASI, 16 AGH). There were two main differences between Group 1 and Group 2. First, we replaced the elevated zero maze with the splash test in Group 2. Second, because we did not initially see any differences in ethanol self-administration in Group 1, we conducted a sucrose fade, as described below, to train the HS rats to drink ethanol. Group 2 was then given the sucrose fade immediately after behavior, followed by increasing concentrations of alcohol. An overview of the study timelines is shown in Fig. 1.

Fig. 1.

Timelines outlining study 1 and study 2 experimental designs. Thirty-two male rats were included in each study. Rats were placed in their assigned housing condition at 4 weeks of age. Weeks relative to housing condition are given.

Behavioral testing

Elevated plus maze (EPM)

Rats were placed at the junction of two elevated, perpendicular arms, one axis with walls and one axis without walls (axis length 111.8 cm, wall height 74.0 cm, floor to apparatus 45.7 cm). Room lights were turned off and the maze was illuminated from above by 15 W incandescent lamps at the extreme end of each open arm (10–12 lux at the extreme ends of open arms). Time spent in closed or open arms was measured over 5 min via beam break at arm entrances (Med-PC IV, Med Associates, Inc.; Fairfax, Vermont, United States). Anxiety-like behavior is associated with increased time in the closed arms (Handley & Mithani, 1984; Pellow, Chopin, File, & Briley, 1985).

Elevated zero maze (EZM)

The EZM is a raised circular apparatus in which two quadrants of a circular arena are enclosed by raised arms, while the other two quadrants remain open. Rats were placed at the entrance to a closed quadrant of the elevated zero maze (wall height 20.5 cm, outer diameter 119.4 cm, platform width 10.2 cm, floor to apparatus 63.5 cm). Room lights were turned off and the maze was illuminated from above. Time spent in the closed or open quadrants was measured over 5 min (Ethovision, Noldus; Leesburg, Virginia, United States). Anxiety-like behavior is associated with increased time in the closed quadrants (Kulkarni, Singh, & Bishnoi, 2007).

Open field test (OFT)

Rats were placed in an empty rectangular arena (40.8 × 40.8 × 30.1 cm) and allowed to explore for 30 min. Movement was monitored via beam break and analyzed via software (Omnitech Electronics, Inc.; Columbus, Ohio, United States) for total distance, movement and rest episodes, rearing episodes, center time, and stereotypic activity. ASI-induced anxiety-like behavior is associated with hyperlocomotion and increased center time (Chappell et al., 2013; Gentsch, Lichtsteiner, & Feer, 1981; Varty, Marsden, & Higgins, 1999).

Adolescent forced swim test (FST)

Rats were placed in a tank of water (25 ± 2 °C, diameter 27.94 cm, height 40.6 cm, water depth 33.0 cm) for 6 min, and the last 5 min of this test were scored manually via video playback for swimming, climbing, and immobility behaviors as described below. The traditional test, as described below, uses a 2-day test with a 15-min swim on day 1. Because a 15-min swim can be stressful on adolescent rats, we used the 1-day paradigm. Previous work (Holl et al., 2018) has shown high correlation between the day 1 immobility and the last 5 min of day 2 immobility in adolescent rats. Scoring was conducted as described below.

Adult FST

Rats were placed in a tank of water (25 ± 2 °C, diameter 28.8 cm, height 49.8 cm, water depth 39 cm) for 15 min on day 1 and 5 min on day 2 as previously described (Solberg et al., 2004). Video recording of day 2 was used to manually score movements made by the rat at 5-s intervals: immobility, swimming, climbing, and diving. Immobility was defined as general lack of movement beyond small movements of a leg or the tail for stabilization. Depressive-like and/or passive coping behavior is associated with increased immobility (Anyan & Amir, 2018; Commons, Cholanians, Babb, & Ehlinger, 2017; Porsolt, Bertin, & Jalfre, 1977; Sharma & Fulton, 2013; Soulis, Papalexi, Kittas, & Kitraki, 2007).

Splash test (SpT)

Rats were placed individually within their home cage and sprayed with a 10% sucrose solution on their dorsal coat. Latency to groom and time spent grooming were measured over 5 min. Higher latency and shorter total grooming time is reported as depressive-like behavior based on a reduced interest in self-care (Ducottet & Belzung, 2004; Mutlu et al., 2012).

Ethanol self-administration

After behavioral testing was complete, when rats were 14 weeks of age, all rats were placed into individual cages to measure voluntary ethanol self-administration, as previously described (Chappell et al., 2013). Ethanol was presented in the home cage at 9:00 AM as a standard 2-bottle choice (bottle order balanced on a per day basis, i.e., all cages received the same bottle order on a given day and the order was switched between days) and administered intermittently (3 × per week) to encourage consumption (Simms et al., 2008). Self-administration was measured after 30 min (binge-like intake, 9:00 AM to 9:30 AM) and after 24 h. To test HS rats’ propensity to self-administer ethanol, Group 1 was initially provided 20% ethanol solution alongside water. We initially found that HS rats drank very small amounts of the 20% ethanol solution. Palatability of ethanol has been shown to be strain-dependent and is thus likely to have contributing genetic factors (Phillips, Crabbe, Metten, & Belknap, 1994; Yoneyama, Crabbe, Ford, Murillo, & Finn, 2008). To encourage ethanol self-administration and achieve greater variability across both groups, we initiated a sucrose fade to improve ethanol palatability (Samson, 1986). The sucrose fade design is outlined in Table 1. Each day for 1 h, an ethanol/sucrose ratio was provided ad libitum alongside water in the home cage. The next ethanol/sucrose ratio was provided the following day. We started with a concentration of 2% ethanol/7% sucrose and ended with a concentration of 8% ethanol/1% sucrose. Ethanol self-administration at 8% ethanol/1% sucrose was then measured in Group 1 for binge-like intake (30 min) and 24-h consumption. Rats in study 2 began ethanol exposure with the sucrose fade and then were given increasing concentrations of ethanol to assess differences in self-administration response at higher concentrations, as well as ethanol intake plateau, between AGH and ASI animals. Ethanol self-administration in study 2 was measured after sucrose fade procedures gradually increased ethanol concentration to 20% ethanol/1% sucrose and then 10% ethanol increases up to 70% ethanol/0% sucrose (Table 2).

Table 1.

Sucrose fade procedure for study 1.

%Ethanol (v/v)	%Sucrose (w/v)
2	7
2	5
4	5
4	3
4	3
5	2
5	1
6	2
6	1
7	2
8	1

Table 2.

Sucrose fade procedure and incremental ethanol increase for study 2.

%Ethanol (v/v)	%Sucrose (w/v)
0	10
2	7
2	5
4	5
4	3
5	2
5	1
6	2
6	1
7	2
8	1
9	1
10	1
8	1
10	1
11	1
12	1
13	1
14	1
15	1
16	1
17	1
18	1
19	1
20	1
30	1
40	1
50	1
60	1
70	0
60	1

Necropsy

Animals were euthanized by decapitation in accordance with IACUC approved methods. Brains were harvested, flash frozen in liquid nitrogen, and stored at −80 °C. Select brains from study 2 were processed to allow determination of protein concentration of GluA1 and SK2 using Western blots.

Western blot analysis

Protein was isolated from dorsal or ventral hippocampi (vHC) from ASI rats (n = 6 and 7, respectively) and AGH rats (n = 7 and 6, respectively). Animals were chosen to represent the higher and lower ethanol self-administration individuals from each group. dHC and vHC were homogenized in homogenization buffer (50 mM Tris-Base, pH 7.35; 20 mM HEPES, pH 7.4; 5 μM EDTA, pH 8.0; protease and phosphatase inhibitor cocktail [Halt™, Thermo-Fisher]). Synaptoneurosomes were prepared as previously described in Sosanya et al., 20l3. Samples were dissolved in SDS sample buffer and resolved on a 4–20% gradient SDS-polyacrylamide gel (Bio-Rad) and transferred to nitrocellulose for western blot analysis (Sosanya et al., 20l3). Antibodies used for western blot analysis were rabbit anti-SK2 (1:500; Alomone Lab; Jerusalem, Israel), mouse anti-actin (1:10 000; Sigma; St. Louis, Missouri, United States), and fluorescent secondary antibodies (AF680; AF800; 1:4000; LiCor; Lincoln, Nebraska, United States). Blots were imaged using LICOR Odyssey CLx infrared imaging system, and densitometry analyses were performed using ImageJ (National Institutes of Health) software. A similar analysis was conducted to measure SK2 expression in vHC synaptosomes from a cohort of three ASI and three AGH Long-Evans rats.

Statistical analysis

All phenotypes were transformed to closely reflect a normal distribution within each housing group prior to analysis. To evaluate the effect of housing on behavioral tests (EPM, EZM, OFT, SpT, and FST), we used lme4 (Bates, Mächler, Bolker, & Walker, 2015) to build linear mixed-effects models of the form: Phenotype ~ housing + sibling status + error, in which “housing” is a fixed effect and sibling status is a random effect to account for the complex family relationships of the HS. Ethanol self-administration was measured in study 1 via a repeated-measures ANOVA at 20% ethanol and over a single 24-h time point following the sucrose fade, and analyzed for differences between ASI and AGH via the linear mixed-effects model above. In study 2, ethanol self-administration was assessed at the single 24-h time point following the sucrose fade using the mixed model above. In addition, all ethanol self-administration sessions from 20% ethanol/1% sucrose to 70% ethanol/1% sucrose were analyzed via repeated-measures ANOVA in GraphPad Prism 8 (GraphPad Software; San Diego, California, United States).

Correlations within and between ethanol self-administration and behavioral outcomes were analyzed via Pearson correlation with Benjamini-Hochberg procedure for multiple comparisons.

Results

High variability in adolescent immobility in the FST

To investigate the potential of the FST to predict behavioral outcomes in adulthood, we measured baseline FST behavior at 3 weeks of age (after weaning and prior to housing assignment). As expected (Holl et al., 2018), we see a large degree of variation in immobility, climbing, and swimming in the adolescent FST (Fig. 2). There were no differences between rats assigned to ASI and AGH for immobility [Fig. 2a, F(1,62) = 0.1346, p = 0.7153], climbing [Fig. 2b, F (1,62) = 2.7679, p = 0.1028], or swimming [Fig. 2c, F (1,62) = 0.7378, p = 0.3947].

Fig. 2.

Baseline and post-ASI behavior in the forced swim test. Prior to housing assignment, no differences were found between AGH (n = 32) and ASI (n = 32) groups in the FST at 3.5 weeks of age, including (a–c) immobility, climbing, and swimming. FST behaviors measured after 8–9 weeks of AGH (n = 32) or ASI (n = 32) show that ASI (d–e) decreased immobility and increased climbing, but (f) no differences were found in swimming. (g) ASI rats showed increased total movement (swimming + climbing + diving) in the FST compared to AGH. Mean ± standard error shown. Open circles represent individual HS rats assigned to AGH, while open triangles represent individual HS rats assigned to ASI housing conditions. *p < 0.05.

ASI induces OFT hyperactivity

We found increased hyperactivity in the OFT in ASI rats compared to AGH rats. Specifically, ASI rats exhibited significantly increased total distance and movement time, and decreased rest time during the first 5 min of the test [Fig. 3a–c, F (1,62) = 11.75, p = 0.0013; F (1,62) = 9.73, p = 0.0031; F (1,62) = 14.76, p = 0.0004, respectively], as well as during the full 30 min of the test [Fig. 3e–g, F (1,62) = 14.422, p = 0.0004; F (1,62) = 19.228, p = 0.0000655; F (1,62) = 19.228, p = 0.0000655, respectively].

Fig. 3.

Open field test behavior. ASI induced hyperactivity in OFT after both 5 min (a–d) and 30 min (e–g) of testing, with no differences found in center time after 5 min (e). Mean ± standard error shown. Open circles represent individual HS rats assigned to AGH, while open triangles represent individual HS rats assigned to ASI housing conditions. *p < 0.05, **p < 0.001, ***p < 0.0001.

ASI has no effect on anxiety-like measures

There were no differences between ASI and AGH for OFT center time during the first 5 min [Fig. 3d, F (1,62) = 2.9696, p = 0.0914] or the full 30 min [Table 3, F (1,62) = 1.2865, p = 0.2624]. In addition, no differences were found in the EPM closed arm or open arm time [Table 3, F (1,62) = 0.0339, p = 0.8546; F (1,62) = 0.4071, p = 0.5265, respectively] or EZM closed area or open area time [Table 3, F (1,30) = 0.8379, p = 0.3695; F (1,30) = 0.8265, p = 0.3727, respectively]. Unlike OFT results, no difference was found between ASI and AGH in EZM total distance traveled [F (1,30) = 1.6924, p = 0.2062, data not shown] or EPM closed arm entries, a measure of overall locomotor activity [F (1,62) = 1.1521, p = 0.2884, data not shown].

Table 3.

No effect of ASI on additional affective phenotypes. Data presented are means (±SD).

Assay	Phenotype	AGH	ASI
EPM	Open Arm Time (s)	84.7 (±41.8)	78.2 (±40.1)
	Closed Arm Time (s)	213.6 (±41.2)	216.1 (±37.2)
EZM	Open Area Time (s)	75.3 (±49.6)	88.7 (±42.5)
	Closed Area Time (s)	223.0 (±49.9)	209.5 (±43.3)
OFT	30 min Center Time (s)	92.93 (±68.96)	75.86 (±45.12)
SpT	Latency to Groom (s)	148.4 (±93.7)	127.3 (±93.6)
	Total Grooming Time (s)	58.9 (±59.5)	60.9 (±39.8)

ASI increases climbing activity in FST

Despair-like/passive-coping behavior was assessed via FST after 8–9 weeks of ASI or AGH. Interestingly, ASI rats showed significantly decreased immobility and increased climbing or escape behavior compared to AGH rats [Fig. 2d–e, F (1,62) = 8.3353, p = 0.0054; F (1,62) = 6.7619, p = 0.0127, respectively). Although no differences were found in swimming between ASI and AGH rats [Fig. 2f, F (1,62) = 1.7897, p = 0.1860], ASI rats did show an overall increase in mobility (swimming, climbing, and diving) compared to AGH rats [Fig. 2g, F (1,62) = 4.6169, p = 0.0356]. Analyses were run with and without including adolescent immobility as a covariate in the linear mixed-effects model. The impact of adolescent FST as a covariate was minimal and did not result in significant differences in the adult data.

No effect of ASI on self-grooming in SpT

The splash test, an additional assay for despair-like behavior, was administered after 9 weeks of ASI or AGH. While we found despair-like/passive coping behavior differences in the FST, no differences were seen in total grooming time [Table 3, F (1,30) = 0.3026, p = 0.5877] or latency to begin grooming [Table 3, F (1,30) = 0.4056, p = 0.5307] between ASI and AGH in the SpT.

ASI decreases self-administration of high ethanol concentrations in HS rats

In study 1, rats were separated into individual cages after 10 weeks of assigned housing and given access to 20% ethanol alongside water for 24 h/day, 3 days/week. There was not a significant difference in ethanol self-administration between ASI and AGH rats over 2 weeks of assessment [Fig. 4a, F (1,29) = 0.2355, p = 0.6312]. To address the possibility that palatability was affecting self-administration in HS rats (Phillips et al., 1994; Yoneyama et al., 2008), we initiated a sucrose fade to introduce ethanol with sucrose (Simms et al., 2008; Wise, 1973). While ethanol self-administration increased across both ASI and AGH by the end of the sucrose fade, there was still no significant difference between the groups in a 24-h assessment of ethanol self-administration [Fig. 4b, F (1,30) = 0.5233, p = 0.4767] or binge-like intake [F (1,30) = 0.0025, p = 0.9605, data not shown].

Fig. 4.

Ethanol self-administration in study 1. (a) Self-administration of 20% ethanol over 14 days was not significantly different between AGH (n = 16) and ASI (n = 16) rats after 10 weeks of assigned housing. (b) There was no difference between AGH and ASI in ethanol self-administration (8% ethanol/1% sucrose) following a sucrose fade procedure. Mean ± standard error shown. Open circles represent (a) mean of AGH rats or (b) individual HS rats assigned to AGH, while open triangles represent (a) mean of ASI rats or (b) individual HS rats assigned to ASI housing conditions.

Based on results in study 1, we began the ethanol self-administration in study 2 with a sucrose fade. Rats were then measured for self-administration at 20% ethanol/1% sucrose, and no difference was seen after 24 h [ Fig. 5a, F (1,30) = 0.7466, p = 0.3964] or for binge-like drinking [F (1,30) = 0.4059, p = 0.5303, data not shown]. We then incrementally increased the concentration of ethanol to investigate tolerance of less palatable concentrations (Table 2). Over the regimen of increased concentrations (from 20% ethanol/1% sucrose to 70% ethanol/1% sucrose), we found a significant interaction between ethanol concentration and housing condition, with AGH rats self-administering higher amounts of ethanol as the concentration increased relative to ASI rats [Fig. 5b, F (34,1020) = 1.609, p = 0.0156]. Throughout study 1 and study 2, we observed high variability in ethanol self-administration (Study 2 shown in Fig. 6), suggesting that genetics may be playing a larger role than environment (e.g., early life stress) in susceptibility to ethanol intake in this population.

Fig. 5.

Ethanol self-administration in study 2. After 10 weeks of AGH (n = 16) or ASI (n = 16), rats underwent a sucrose fade procedure and showed no difference in (a) ethanol self-administration (20% ethanol/1% sucrose). (b) AGH rats self-administered higher amounts of ethanol relative to ASI as ethanol concentration increased over time. Vertical bars represent increases in ethanol concentration ([ethanol]/[sucrose]: 20/1,30/1,40/1,50/1,60/1,70/0,60/1). Mean ± standard error shown. Open circles represent (a) individual HS rats assigned to AGH or (b) mean of AGH rats, while open triangles represent (a) individual HS rats assigned to ASI housing conditions or (b) mean of ASI rats.

Fig. 6.

Individual variation in ethanol self-administration over time. Self-administration of ethanol for individual rats is shown for (left) AGH and (right) ASI in study 2. Vertical bars represent increases in ethanol concentration ([ethanol]/[sucrose]: 20/1, 30/1, 40/1, 50/1, 60/1, 70/0, 60/1). Each color represents an individual animal.

Adolescent FST correlates with adult EPM but not adult FST

Surprisingly, we did not find correlations between adolescent FST and adult FST measures (immobility, r = −0.0383, p = 0.7639; climbing, r = −0.1412, p = 0.2659; swimming, r = −0.0870, p = 0.4944; data not shown). However, adolescent climbing behavior in the FST strongly correlates with OFT center time, although only in ASI rats (Fig. 7, ASI: r = 0.4716, p = 0.0064; AGH: r = 0.0111, p = 0.95). These data indicate that active coping in adolescence may predict response to early life stress, with these animals showing a lower anxiety-like response in adulthood. Additional behavioral correlations can be found in supplementary material.

Fig. 7.

Adolescent forced swim test climbing correlates with adult open field test center time in ASI rats. FST behavior measured prior to AGH (n = 32) or ASI (n = 32) showed a positive correlation between FST climbing and open field test center time only in ASI rats (ASI: r = 0.4716, p = 0.0064; AGH: r = 0.0111, p = 0.95). Open circles represent individual HS rats assigned to AGH, while open triangles represent individual HS rats assigned to ASI housing conditions.

ASI has no effect on the expression of SK2 or GluA1 in synaptosomes isolated from either the dorsal or ventral hippocampus in HS rats

Concordant with the lack of effect of ASI on anxiety-like behaviors and ethanol self-administration, we found no effect of ASI on SK2 expression in the synaptosomal fraction of either the dHC or vHC [Fig. 8a–d, GluA1 vHC F (1,11) = 5.3739, p = 0.0604; SK2 vHC F (1,11) = 1.2881, p = 0.2805; GluA1 dHC F (1,11) = 0.0043, p = 0.9502; SK2 dHC F (1,11) = 0.0744, p = 0.7901]. We do note that ASI rats exhibit a trend toward decreased expression of GluA1 in the vHC [F (1,11) = 5.3739, p = 0.0604], with no difference in the dHC [F (1,11) = 0.0043, p = 0.9502]. Because Long-Evans rats do show the expected increase in anxiety-like behavior and alcohol consumption in response to ASI (Chappell et al., 2013), we also examined SK2 expression in vHC synaptosomes from ASI and AGH Long-Evans rats. As expected, in these Long-Evans cohorts, we observed a significant ASI-mediated decrease in SK2 expression [T (1,5) = 3.53, p = 0.0241, data not shown].

Fig. 8.

GluA1 and SK2 levels in the ventral and dorsal hippocampus of HS rats in ASI or AGH conditions. No differences were found between GluA1 or SK2 protein levels in vHC or dHC. Representative Western blots are shown of GluA1 and SK2 protein levels in (a) vHC and (b) dHC. Western blot analysis and quantification in study 2 showed no difference in protein levels of GluA1 or SK2 in the (c) vHC (AGH n = 6 and ASI n = 7) or (d) dHC (AGH n = 7 and ASI n = 6). Mean ± standard error shown. Open circles represent individual HS rats assigned to AGH, while open triangles represent individual HS rats assigned to ASI housing conditions.

Discussion

Numerous prior studies have demonstrated that in many rodent strains, ASI leads to significant increases in locomotor activity and negative affective behaviors, including anxiety-like measures. These behaviors can be considered risk factors for AUD and, consistent with that notion, ASI often promotes long-lasting increases in operant and home-cage measures of ethanol self-administration. Here, we report that HS rats, an outbred rat model used for genetic mapping, exhibit hyperactivity in response to ASI, but are not susceptible to ASI-induced increases in anxiety-like behaviors or ethanol self-administration, suggesting that this strain may be resilient to at least some of the maladaptive behavioral phenotypes associated with early life stress. We further show that, while ASI leads to decreases in hippocampal SK2 expression in Long-Evans rats, HS rats do not exhibit changes in the expression of two synaptic proteins, SK2 and GluA1, which likely contribute to the ASI-induced behavioral changes in these other strains. These findings suggest that the lack of plasticity in SK2 and GluA1 levels in the vHC may represent a potential molecular mechanism conferring protection against ASI-induced behavioral changes in HS rats. These data are in stark contrast to our original hypothesis and indicate that instead of being a model to understand how genes interact with early life stress to predispose individuals to anxiety and increased drinking behavior, HS rats are instead a model in which to identify molecular mechanisms that protect against these negative effects.

Similar to previous studies in Long-Evans, Sprague Dawley, and Wistar rats (Chappell et al., 2013; Heidbreder et al., 2000; Reinwald et al., 2018), we demonstrate that ASI leads to behavioral hyperactivity in HS rats. Specifically, we find that HS rats exposed to ASI exhibit increased total distance and movement time in the OFT relative to AGH rats. These results indicate that the HS rat could be used to study genetic interactions underlying susceptibility to ASI-induced hyperactivity. Previous studies have shown that hyperactivity in response to ASI may be a result of basal turnover of serotonin in the nucleus accumbens, which is linked to isolation-induced hyperactive exploratory behavior (Heidbreder et al., 2000). Alternative studies have utilized brain imaging and neural network architecture analyses to correlate these behavioral abnormalities with connectivity alterations resembling those found in neurodevelopmental psychiatric disorders, including attention-deficit/hyperactivity disorder (Reinwald et al., 2018).

In contrast to previous work by us and others using Long-Evans rats (Chappell et al., 2013; Karkhanis et al., 2016; McCool & Chappell, 2009; Yorgason et al., 2013) and other outbred rodent strains reporting robust and enduring increases in anxiety-like behaviors and ethanol self-administration (for review, see Butler et al., 2016), we did not find differences in OFT, EPM, or EZM measures of anxiety-like behavior nor in ethanol self-administration in response to ASI in the HS rat. If anything, after a sucrose fade, we saw a slight but significant decrease in ethanol self-administration in ASI rats relative to AGH rats. Relative to other strains, the HS rat has been shown to be anxious with passive-coping strategies (Díaz-Morán et al., 2013), and it is possible that this phenotype could contribute to altered biochemistry that protects against ASI-induced anxiety and ethanol intake. Additionally, it has been shown that addiction vulnerability behaviors are independent traits in HS rats, counter to other outbred strains where these traits are correlated (Hughson et al., 2019), further indicating that the HS population differs behaviorally from other strains. We purport that the lack of an observed ASI-induced behavioral change in these measures may be due to differences in molecular mechanisms that protect the HS rats from ASI-induced anxiogenic and ethanol drinking behaviors.

To investigate potential neuronal adaptations induced by ASI, we analyzed the expression of the small conductance, calcium-activated potassium channel, SK2, and the GluA1 AMPA receptor subunit in synaptoneurosomes isolated from the dorsal and ventral hippocampus. Activation of SK2 channels strongly buffers synaptic excitation in the hippocampus, and in previous work, we reported that ASI-induced increases in anxiety-like behavior and ethanol consumption in Long-Evans rats are associated with a significant increase in synaptic excitability in the ventral hippocampus (Almonte et al., 2017). We also observed a significant decrease in SK2 expression in ventral hippocampal synaptosomes and an increase in GluA2 expression following withdrawal from a chronic intermittent ethanol vapor exposure regimen that increases measures of anxiety-like behavior and ethanol self-administration (Ewin et al., 2019). SK2 expression is enriched in the ventral hippocampus, and activation of these channels serves to powerfully dampen glutamatergic synaptic excitation. Consistent with these findings, we found that SK2 expression was indeed reduced in ventral hippocampal synaptosomes from Long-Evans rats following ASI. In contrast to these findings, and consistent with the current behavioral studies in HS rats, ASI had no effect on either GluA1 or SK2 expression in ventral hippocampus synaptosomes. These findings may partially explain why ASI failed to induce increased anxiety-like behavior and ethanol self-administration in the HS rat. Future studies will be able to determine whether these two proteins play a causal role in protecting HS rats from ASI-induced increases in anxiety-like behavior and ethanol self-administration.

The lack of ASI-induced increases in ethanol self-administration in HS rats may also be because genetic susceptibility to alcohol intake varies significantly within the HS rat population. Previous work has shown that there is large variation in ethanol self-administration between the HS founder strains (Pohorecky, 1984). In addition, the HS rat has previously been utilized to create high- and low-alcohol drinking rats and to identify candidate genes within QTL (Bice et al., 2010). These results suggest that the genetic susceptibility to alcohol intake inherent within the HS rat population may be sufficient to override the early life stress of ASI on ethanol intake. Indeed, we observed a wide range of drinking behaviors across ASI and AGH. Another possibility is that we were not able to determine the optimal concentration of ethanol solution, with or without sucrose, in this study. Future studies utilizing the HS population in an ethanol self-administration paradigm should further investigate methods to initiate and sustain voluntary ethanol intake.

In addition to anxiety-like behaviors and ethanol self-administration, we used the FST to assess despair-like/passive-coping behavior in response to ASI. Interestingly, we found decreased immobility and increased climbing in ASI rats, suggesting decreased, as opposed to increased, despair-like behavior. Previous studies have shown conflicting results in the FST in ASI models, with some studies showing increased immobility (Brenes, Rodríguez, & Fornaguera, 2008; Kokare, Dandekar, Singru, Gupta, & Subhedar, 2010) and others showing increased climbing (Hong et al., 2012) or no effect (Fischer, Liebenberg, Elfving, Lund, & Wegener, 2012; Hall, Huang, Fong, & Pert, 1998). Recent commentary surrounding the FST has questioned its interpretation as a test of depression-like behavior, instead suggesting that FST phenotypes could be an extension of anxiety-like behavior (Anyan & Amir, 2018) or coping response to stress (Commons et al., 2017; de Kloet & Molendijk, 2016). The authors, Anyan and Amir (2018), point toward reduced activity levels in genetic models of anxiety, such as reduced exploratory behaviors in the EPM and OFT, as a factor contributing to increased FST immobility. In contrast, our results show clear differences in FST behavior, including decreased immobility in response to ASI, without a concomitant alteration of OFT, EPM, or EZM anxiety-like behaviors, indicating that FST is measuring a phenotype separate from EPM and EZM in this model. Similar to our current findings, previous work has shown that ASI female Sprague Dawley rats exhibit increased FST climbing behavior with increased sucrose preference relative to non-ASI rats, although these differences were not seen in Sprague Dawley males (Hong et al., 2012). Although not previously discussed in the literature, an alternative hypothesis is that the increased climbing and decreased immobility in response to ASI in the HS rats represents overall hyperactivity in these animals (perhaps a coping response to stress), and can be seen as an extension of the ASI-induced hyperactivity seen in the OFT.

We also assessed the relationships between adolescent FST and adult behavioral phenotypes in this model. We show that adolescent FST behavior is not predictive of FST phenotype in adulthood. This may indicate that the FST is measuring different behaviors in adolescence vs. adulthood, although repeated testing in the FST may be confounding these results (Bogdanova, Kanekar, D’Anci, & Renshaw, 2013). Interestingly, adolescent FST climbing positively correlated with OFT center time in ASI rats, suggesting that active coping behavior in adolescence may predict decreased anxiogenic response to adolescent stress. This is a novel and potentially very interesting finding.

Limitations of this study include the exclusive use of male rats and relatively low sample sizes in protein expression analysis. Previous work in our lab has shown no effect of ASI on ethanol self-administration in female Long-Evans rats, contrasting with ASI-induced increases in ethanol self-administration in males (Butler, Carter, & Weiner, 2014). Considering the differences found here between Long-Evans males and HS males, it would be interesting to see whether sex differences also exist in the HS response to ASI, or whether HS females are also protected from the negative effects of ASI. Additionally, increasing the number of rats analyzed for GluA1 and SK2 protein expression would provide more power to demonstrate potential differences, or lack thereof, between ASI and AGH groups, especially given the strong trend found in GluA1 vHC expression.

In conclusion, we have shown that the HS population is susceptible to some ASI-induced behavioral changes, including OFT hyperactivity and FST behavior, but not increased anxiety-like behaviors or ethanol self-administration, demonstrating a separation of ASI effects in this model. These behavioral results are mirrored by similar differences in the expression of GluA1 and SK2, with differences seen in Long-Evans rats, but not HS rats. These results indicate that neurobiological alterations that govern glutamatergic excitability may confer protection from ASI-induced anxiety-like and drinking behaviors in the HS rats. Future studies in HS rats will further investigate this relationship as well as identify additional factors that may be contributing to this protection.