Early developmental alcohol exposure alters behavioral outcomes following adolescent re‐exposure in a rat model

Abstract

Background

Prenatal alcohol exposure alters brain development, affecting cognitive, motor, and emotional domains, and potentially leading to greater alcohol intake during adolescence. The present study investigated whether early alcohol exposure modifies vulnerability to behavioral alterations associated with adolescent alcohol exposure in a rodent model.

Methods

Sprague–Dawley rats received ethanol or sham intubations during two developmental periods: (1) the third trimester equivalent of brain development in humans (postnatal days [PD] 4–9) and (2) adolescence (PD 28–42). Both exposures resulted in blood alcohol concentrations around 200 mg/dl. Subjects were tested in the open field (PD 45–48) and on hippocampal and prefrontal cortical (PFC) dependent tasks: the Morris water maze (PD 52–58) and trace fear conditioning (PD 63–64).

Results

Neonatal alcohol exposure reduced forebrain and cerebellar weight, increased open‐field activity, and slowed acquisition of trace fear conditioning. Adolescent alcohol exposure did not disrupt learning or significantly induce gross brain pathology, suggesting that 200 mg/dl/day of ethanol disrupts cognitive development during the 3rd trimester equivalent, but not during adolescence. Interestingly, females exposed to alcohol only during adolescence exhibited an increased conditioned fear response and more rapid habituation of locomotor activity in the open field, suggesting alterations in emotional responding. Moreover, subjects exposed to a combination of neonatal and adolescent alcohol exposure spent significantly more time in the center of the open field chamber than other groups. Similarly, males exposed to the combination exhibited less thigmotaxis in the Morris water maze.

Conclusions

These results indicate that combined exposure to alcohol during these two critical periods reduces anxiety‐related behaviors and/or increases risk taking in a sex‐dependent manner, suggesting that prenatal alcohol exposure may affect risk for emotional consequences of adolescent alcohol exposure.

The current study investigated whether prenatal alcohol exposure increases vulnerability to the behavioral consequences of adolescent alcohol exposure. Early alcohol exposure alone produced hyperactivity and some learning deficits, while adolescent alcohol exposure alone altered emotional responding in females. However, the combination of early and adolescent alcohol led to increases in behaviors indicative of reduced anxiety and/or increased risk‐taking in a sex‐ and task‐dependent manner. Thus, prenatal alcohol exposure combined with adolescent alcohol exposure may specifically increase risk for emotional dysregulation.

INTRODUCTION

Alcohol is a known teratogen that can produce Fetal Alcohol Spectrum Disorders (FASD), a wide range of developmental problems including physical malformations, functional and structural damage to the developing brain, and cognitive and behavioral impairments (Riley et al., 2011; Tsang et al., 2016; Williams et al., 2015). Yet, FASD are still a prevalent public health concern, as an estimated 1–5% of children are born with FASD in the United States (May et al., 2018).

FASD alterations may include increased alcohol preference and alcohol consumption later in life, and may also increase the risk of developing alcohol use disorders (AUD) (Cornelius et al., 2016; Fabio et al., 2015; Goldschmidt et al., 2019). Moreover, early alcohol exposure may make the brain more vulnerable to the effects of alcohol during adolescence (Maldonado‐Devincci et al., 2010). In fact, adolescent alcohol exposure on its own increases chances of AUD as the prefrontal cortex (PFC), hippocampus, and other neural systems are still developing (Baer et al., 2003; Bava & Tapert, 2010; Doremus‐Fitzwater et al., 2010; Witt, 2010). Adolescence is a particularly vulnerable period during which alcohol poses risks to development and causes greater deficits compared to the effects of alcohol exposure during adulthood (Crews et al., 2016; Seemiller & Gould, 2020; Spear, 2018). This is concerning given that a significant portion of the adolescent population in the US engage in binge drinking (Gaztañaga et al., 2020; Johnston et al., 2017), defined as achieving a blood alcohol concentration of 0.08% or above (NIAAA, 2022). Earlier onset of drinking doubles the risk of developing AUDs (Addolorato et al., 2018; Brown et al., 2008). However, it is unclear how a history of prenatal alcohol exposure impacts behavioral outcomes for those who also drink alcohol again during adolescence.

In humans, repeated binge‐like alcohol consumption causes more severe memory impairments in adolescents when compared to adults (Witt, 2010), leads to long‐term changes in neurotransmitter systems, results in poorer neurocognitive performance (Seemiller & Gould, 2020), and damages white matter in the hippocampus and PFC (Jacobus & Tapert, 2013; McQueeny et al., 2009; Spear, 2018). Similar findings are also seen with animal models (Nixon et al., 2010; Nixon & McClain, 2010). In rodents, adolescent binge‐like alcohol exposure alters hippocampus and PFC function via widespread impact on neurogenesis, neural connectivity, and receptor function (Ehlers et al., 2013; Morris et al., 2010; Pascual et al., 2007).

Neuropathology associated with adolescent alcohol exposure may lead to long‐lasting behavioral alterations, including deficits in spatial learning (Sircar & Sircar, 2005). For example, rats exposed to intermittent adolescent alcohol binges show decreased memory retention and accelerated forgetting in the Morris Water Maze (Schulteis et al., 2008). Rats exposed to binge‐like alcohol during adolescence also show hypoactivation on locomotor tasks (Ehlers et al., 2013), lowered anxiety‐related behaviors in an open field during adulthood (Ehlers et al., 2013), and long‐term deficits in conditional discrimination learning and object recognition tasks; these deficits are consistent with alcohol‐induced hippocampal damage (Pascual et al., 2007).

The PFC also undergoes changes during adolescence and is especially vulnerable to damage caused by binge drinking (Crews & Boettiger, 2009; Oscar‐Berman & Marinković, 2007). Importantly, adolescents with an AUD have smaller overall PFC and PFC white matter volumes than controls (De Bellis et al., 2005; Medina et al., 2008). Human adolescents with a history of binge drinking also show impairments on tests of executive functioning (Parada et al., 2012) as well as sustained attention tasks and memory recall (Hartley et al., 2004). Similarly, rats exposed to binge‐level intermittent alcohol exhibit deficits in reversal learning in the Barnes maze task, indicative of PFC damage (Vetreno & Crews, 2012). Animals exposed to alcohol in a binge‐like manner also show impairments in both contextual fear conditioning (hippocampus‐dependent) and trace fear conditioning (hippocampus and PFC dependent), but not in delay conditioning, showing that the hippocampus and PFC in particular may be especially sensitive to an alcohol insult during adolescence (Murawski & Stanton, 2011; Seemiller & Gould, 2020; Spear, 2018; Yttri et al., 2004). Further, acute binge‐like alcohol exposure disrupts trace conditioning in adolescent rodents, but not adults. Conversely, binge‐like alcohol impaired contextual learning in adults, but not adolescents, highlighting age‐based differential susceptibility to acute alcohol effects (Hunt & Barnet, 2016).

However, it is not clear how alcohol exposure during both developmental periods (prenatal and adolescence) affects behavior differently than exposure during each period alone. The present study uses an animal model to examine whether early alcohol exposure modifies the effects of adolescent alcohol exposure on behaviors that depend on the functional integrity of the hippocampus and PFC. In other words, do rats that are exposed to alcohol in adolescence after previous exposure during early development perform worse on behavioral tasks in adulthood than rats that are exposed to alcohol during adolescence without a prior history of exposure? Early exposure to alcohol may make the brain more vulnerable to damaging effects of alcohol during subsequent exposures (Maldonado‐Devincci et al., 2010), which could have important implications for adolescents who binge drink.

To address this question, the present study used a model of alcohol exposure during the 3rd trimester brain growth spurt, a period of brain development that occurs postnatally in rats (herein referred to as the neonatal period), during which the brain is particularly vulnerable (Goodlett & Johnson, 1997). Subjects were exposed to alcohol during the neonatal period alone, the adolescent period alone, or the combination. In early adulthood, subjects were tested in (1) an open field, (2) Morris water maze spatial learning task and (3) context and trace fear conditioning.

MATERIALS AND METHODS

The present study included a total of 90 Sprague–Dawley rats in a 2 (neonatal alcohol exposure, control) × 2 (adolescent alcohol exposure, control) × 2 (male, female) design. Subjects were assigned to 1 of 4 exposure groups: (1) neonatal alcohol exposure only (n = 21; 10 females, 11 males), (2) adolescent alcohol exposure only (n = 24; 12 females, 12 males), (3) combination of neonatal and adolescent alcohol exposure (n = 21; 11 females, 10 males) and (4) controls (n = 24, 12 females, 12 males).

All litters were generated at the animal facilities at the Center for Behavioral Teratology. Pairs of adult females and males were placed in separate cages overnight. Seminal plugs indicated pregnancy and gestational day (GD) 0. The pregnant females were each singly housed until pups were born, typically GD 22 (which was also denoted as postnatal day [PD] 0). On PD 1, litters were culled to eight animals and subjects were randomly assigned to treatment groups, resulting in one subject per sex per treatment group in each litter. Subjects that received binge‐like neonatal alcohol exposure were treated daily from PD 4–9, a period of brain development that parallels the third trimester period of development in humans. Alcohol was administered via intragastric intubation as 2.5 g/kg/day ethanol (EtOH) (11.39% v/v ethanol concentration in a milk formula with an administration volume of 0.0275 ml/g body weight), followed by one milk feeding 2 h later (Goodlett & Johnson, 1997; Johnson & Goodlett, 2002; Murawski & Stanton, 2011; Ryan et al., 2008). The follow‐up milk feeding was provided, as is standard in the field, because intoxicated pups may suckle less (Johnson & Goodlett, 2002). Control subjects received sham intubations. The neonatal alcohol dose was expected to result in blood alcohol concentrations between 200 and 250 mg/dl. Between intubations, pups remained with the dam. On PD 7, each pup was given a paw tattoo, indicating pup identification number and allowing investigators to remain blind to treatment condition during behavioral testing.

Subjects that received a subsequent binge‐like alcohol exposure during adolescence (PD 28–42) were administered 4.0 g/kg/day ethanol on a 2 day on/1 day off schedule throughout adolescence (18.23% v/v ethanol concentration in water with an administration volume of 0.0275 ml/g body weight) via intubation (McClain et al., 2011; Vetreno et al., 2014; Vore et al., 2017). Control subjects received sham intubations of water. The adolescent alcohol dose was also expected to result in blood alcohol concentrations between 200 and 250 mg/dl.

Blood alcohol concentrations

To determine peak blood alcohol levels (mg/dl), 20 μl of blood were collected via a tail clip 1.5 h after the ethanol treatment on PD 6 during the neonatal exposure period and 1.5 h after adolescent ethanol exposure on PD 35. Blood samples were collected from all the subjects, but samples from sham‐intubated subjects were discarded. Peak BAC was determined using an Analox analyzer (4Model AMI, Analox Instruments; Lunenburg, MA).

Behavioral testing

Following adolescent alcohol exposure, all subjects were tested on a behavioral battery that included open‐field activity, visuospatial learning, and trace fear conditioning. The order of testing was identical for all subjects to minimize stress carry‐over effects between tasks.

Open‐field activity

Starting on PD 45, activity levels were measured in an open field during the dark cycle for 60 min per night (beginning at 06:00 pm) for four consecutive nights (PD 45–48). Activity levels were measured in dark, enclosed, and ventilated open‐field chambers (40 × 40 × 30.5) via an optical beam activity monitor (Hamilton‐Kinder, San Diego, CA). Subjects were transported from the colony room to the activity testing room in the dark, illuminated only by a dim red light. Chambers were cleaned prior to each testing period and white noise was played in the room to minimize odor and sound stimuli, respectively. After 30 min of acclimation to the testing room, each subject was placed in the center of the open field and infrared beam interruptions were recorded. Activity measures were collected in 5‐min sessions over the 60‐min test on each testing day. Total distance traveled, movement time, rearing, center entries and time, and thigmotaxis were examined.

Visuospatial learning

Starting on PD 52, subjects were trained on a hippocampal‐dependent spatial learning task, the Morris water maze (PD 52–58). Subjects were placed in a water tank 1.5 m in diameter (water at 26 °C) in a room equipped with video recording using Water 2020 software (HVS image, San Francisco). During the acquisition phase, subjects used visuospatial cues in the room to find a circular escape platform (4‐in diameter) that was hidden under the water's surface. The location of the platform remained the same across training for each subject. Each subject was trained for 4 trials per day (from different start positions) with an inter‐trial interval of 4–5 min for 6 consecutive days. Each trial lasted no more than 60 s. If the subject did not find the platform in 60 s, they were led to the platform by the experimenter. Subjects were given 10 s on the platform after each trial to study its spatial location. Path‐length (distance to platform; m), latency (time to platform; s), swimming speed (m/s), heading angle (accuracy in orientation toward the platform; degrees), and thigmotaxis (percent time spent along the tank wall) were measured.

On day 7, subjects were tested for memory of platform location using a probe trial. The escape platform was removed from the tank and each subject was placed in the water tank for one 60‐s trial. The number of passes through the target quadrant and amount of time spent in the target quadrant and in the target area (the area 3 times the platform diameter where the platform used to be) were measured as an assessment of memory. After probe testing, subjects underwent trials with a visible platform for 2 days to test for performance variables. During this testing phase, the platform protruded above the water and was made visible, and visual cues around the room were covered. Each subject was tested for 4 trials per day from the same start position, but the platform position changed for each trial (the order of platform rotation remained the same between days).

Trace fear conditioning

Both context and trace fear conditioning occurred over a period of 2 days (PD 63–64). On day 1, subjects were trained to associate a conditioned stimulus (CS), a 10‐s, 80‐decibel white noise, with an unconditioned stimulus (US), a 2‐s 0.6 mA foot shock. The CS and US were separated by a 10‐s trace interval. Training occurred in a novel context (Context A). Following a 120‐s baseline period, subjects experienced 5 CS–US pairings, with an average inter‐trial interval of 120 s. On Day 2, subjects were returned to Context A and tested for freezing over a 300‐s stimulus free test. This measure served as an indication of the contextual learning. Two hours later, subjects were placed in an alternate context (Context B) to test freezing specific to the CS. Onset of the CS occurred after a 120‐s baseline period. The 10‐s CS was followed by 120 s of stimulus free testing (trace and inter trial interval [ITI]), and then by a 2‐min continuous CS. This paradigm measured freezing during the baseline period in context B, freezing at onset of CS, freezing to the trace period after the CS, and freezing to the CS itself.

Brain collection

All the animals were euthanized on PD 65 via CO₂ exposure followed by rapid brain tissue collection. Whole brain, forebrain, and cerebellar weights were recorded. Fresh brain tissues were immediately frozen on dry ice and stored at −80°C for potential later processing.

Data analyses

SPSS software was used to analyze all body weight, blood alcohol level, and behavioral data. Data were analyzed with three between‐subjects factors: neonatal alcohol exposure (NeoEtOH or Sham), adolescent alcohol exposure (AdoEtOH or Sham), and sex (female or male). For open field activity, day and 5‐min bin served as within‐subject repeated measures; for Morris water maze, days and trials served as within‐subject repeated measures. A univariate ANOVA was used to analyze probe day trials of the Morris Water Maze task. For fear conditioning, 5 training trials on day 1 served as within‐subject repeated measures. Context testing data were analyzed using repeated measures ANOVA, and CS Test data were analyzed using a univariate ANOVA. Effect sizes are also reported (η²) for all significant results.

RESULTS

Body weights

There were no significant effects of neonatal alcohol exposure on body growth. Similarly, there were no effects of adolescent alcohol exposure, although females exposed to alcohol lagged in body weight, with significantly lower body weights on the last day of testing only (PD 41; F (1, 43) = 3.96, p = 0.05, η² = 0.08), an effect driven by two female ethanol‐exposed subjects (data not shown).

Blood alcohol concentrations

Blood alcohol concentrations did not differ between the neonatal (PD 6) and adolescent (PD 35) exposure periods. Importantly, a history of neonatal EtOH exposure did not significantly affect BACs during adolescence (Table 1).

TABLE 1.

BACs did not differ between male and female offspring during either exposure period. A history of neonatal EtOH exposure also did not alter BACs reached during adolescence.

Exposure group	n	BAC means ± SEM (mg/dl)
		PD 6	PD 35
Neonatal EtOH + Adolescent EtOH
Females	10	208.38 ± 12.28	192.62 ± 12.09
Males	11	202.75 ± 14.78	202.36 ± 10.06
Neonatal EtOH + Sham
Females	9	198.02 ± 5.21	–
Males	10	206.94 ± 10.60	–
Sham + Adolescent EtOH
Females	12	–	204.65 ± 12.44
Males	12	–	209.52 ± 9.79

Open‐field activity

Locomotor activity decreased within and between sessions, as all subjects habituated to the open field chamber. However, both Neonatal and Adolescent EtOH exposure altered locomotor activity in a sex‐ and time‐dependent manner, producing a 5‐way interaction of Day × Bin × Sex × NeoEtOH × AdoEtOH (F (33, 2673) = 1.8, p < 0.05, η² = 0.02) on the total distance traveled. Follow‐up analyses indicated that females exposed to adolescent EtOH exposure alone (Sham+AdoEtOH) had lower locomotor activity levels compared to those in the combined exposure group (NeoEtOH + AdoEtOH) on each Day (PD 45: F (1, 20) = 4.6, p < 0.05, η² = 0.18; PD 46: F (1, 20) = 6.4, p < 0.05, η² = 0.25; PD 47: F (1, 20) = 4.5, p < 0.05, η² = 0.18; PD 48: F (1, 20) = 5.7, p < 0.05, η² = 0.22; Figure 1). However, neither developmental EtOH exposure nor adolescent EtOH exposure affected activity levels among males (Figure 1). Similar effects were observed in basic and fine movements in the chamber (data not shown). Interestingly, only females exposed to neonatal EtOH, alone or in combination with adolescent exposure, exhibited significant increases in rearing, an exploratory behavior (Sex × NeoEtOH interaction: F (1, 81) = 4.11, p < 0.05, η² = 0.05; Females NeoEtOH: 802.0 ± 58.0, Females NeoSham: 661.4 ± 27.5, Males NeoEtOH: 595.4 ± 38.5, Males NeoSham: 615.0 ± 34.2).

FIGURE 1.

Although neonatal EtOH exposure did not alter overall activity levels, adolescent EtOH exposure significantly reduced locomotor activity among females (A) but not males (B). **Sham + adolescent EtOH < neonatal EtOH + adolescent EtOH, p < 0.05.

In contrast, there were opposing effects of neonatal and adolescent EtOH exposure on locomotor activity in the center of the chamber, and these effects differed by sex, as indicated by interactions of Day × Bin × NeoEtOH × AdoEtOH × Sex (F (33, 2673) = 1.4, p = 0.05, η² = 0.02), Bin × NeoEtOH × AdoEtOH × Sex (F (11, 2673) = 2.2, p < 0.05, η² = 0.03), Bin × AdoEtOH (F (11, 2673) = 2.4, p < 0.01, η² = 0.03), and Bin × Sex (F (11, 2673) = 2.2, p < 0.05, η² = 0.03). Overall, the subjects exposed to EtOH during neonatal development traveled significantly greater distances in the center of the chamber compared to their Sham counterparts, as indicated by a main effect of NeoEtOH (F (1, 81) = 7.3, p < 0.01, η² = 0.08; Figure 2A). However, interactive effects of neonatal and adolescent alcohol were observed in the females (Bin × NeoEtOH × AdoEtOH: F (11, 440) = 1.84, p < 0.05, η² = 0.04). Females exposed to EtOH during the neonatal period alone or in combination with adolescent EtOH exposure traveled greater distances in the center of the chamber (F (1, 40) = 11.64, p < 0.01, η² = 0.23; Figure 2B). In contrast, females exposed to EtOH during the adolescent period alone exhibited reduced activity, primarily during the first half of the testing sessions (first 6 Time Bins; F (1, 20) = 10.5, p < 0.01, η² = 0.21). These interactions were not observed among males (Figure 2C).

FIGURE 2.

Neonatal EtOH exposure increased the distance traveled in the center of the chamber (A), an effect primarily driven by the female subjects. Although neonatal EtOH exposure increased the center distance traveled, adolescent EtOH exposure alone decreased this measure (B), effects not observed among males (C). *Neonatal EtOH > no neonatal EtOH, p < 0.01. **Adolescent EtOH < all other groups, p < 0.01.

Changes in locomotor activity in the center of the chamber may reflect alterations in anxiety‐related behaviors, as rats typically prefer spending time near the walls within the chamber in contrast to the open center area. Subjects exposed to neonatal EtOH exposure, regardless of sex, spent significantly more time in the center of the chamber compared with Sham‐intubated subjects (F (1, 81) = 13.7, p < 0.001, η² = 0.15). However, a 2‐way Bin × AdoEtOH interaction was also present (F (11, 891) = 2.4, p < 0.01, η² = 0.03). At the beginning of the sessions, subjects exposed to EtOH during the neonatal period spent more time in the center than subjects exposed to EtOH during adolescence (F (1, 42) = 8.2, p < 0.01. η² = 0.16). In contrast, subjects exposed to the combination of EtOH during the neonatal and adolescent periods exhibited less habituation, spending more time in the center of the chamber by the midpoint of the session (30 min) (p's < 0.05; Figure 3A). Importantly, only NeoEtOH exposure, alone or in combination with adolescent EtOH exposure, increased the number of entries into the center of the chamber (F (1, 81) = 5.2, p < 0.05, η² = 0.06; Figure 3B).

FIGURE 3.

At the beginning of the activity sessions, subjects exposed to EtOH during the neonatal period spent more time in the center of the chamber; however, the combination exposure group spent more time in the center beginning in the middle of the session (A). Overall, neonatal EtOH exposure increased the number of entries into the center of the chamber (B). **Neonatal EtOH + adolescent EtOH > all other groups, p < 0.05. *Neonatal EtOH > no neonatal EtOH, p < 0.05.

Visuospatial learning

Performance on the Morris water maze improved over training days, producing main effects of Day (F (5, 410) = 113.1, p < 0.01, η² = 0.58), Trial (F (3, 246) = 44.5, p < 0.01, η² = 0.35), and a Day × Trial interaction (F (15, 1230) = 2.1, p < 0.01, η² = 0.03) on path length (Figure 4). Male subjects had shorter path lengths to platform compared with female subjects, producing a main effect of Sex (F (1, 82) = 11.8, p < 0.01, η² = 0.13).

FIGURE 4.

Among females, neither EtOH exposure paradigm altered path lengths during acquisition (A). Neonatal EtOH exposure increased path lengths among males, although not significantly (B).

Heading angle serves as a measure of accuracy in orientation toward the escape platform (Figure 5A,B; Chernyuk et al., 2021; Morris, 1984). A significant interaction of NeoEtOH × AdoEtOH × Sex (F (1, 82) = 5.7, p < 0.05, η² = 0.07), as well as a Trial × NeoEtOH × Sex interaction (F (3, 246) = 2.8, p < 0.05, η² = 0.03) and a main effect of AdoEtOH (F (1, 82) = 4.8, p < 0.05, η² = 0.06) were observed. Follow‐up analyses were conducted for male and female subjects separately to explore these interactions. Unlike open field behavior, no significant effects of EtOH exposure were seen among female subjects. In contrast, interactions of NeoEtOH × AdoEtOH (F (1, 41) = 7.7, p < 0.01, η² = 0.16) and Trial*NeoEtOH (F (3, 123) = 2.9, p < 0.05, η² = 0.07) and a main effect of AdoEtOH (F (1, 41) = 4.5, p < 0.05, η² = 0.10) were observed among males. Male subjects exposed to the combination of neonatal and adolescent EtOH had significantly smaller heading angles compared to those exposed to neonatal EtOH alone (F (1, 19) = 11.5, p < 0.01, η² = 0.38). However, none of the EtOH groups differed significantly from controls. Overall, heading angle declined during acquisition in males, but not females, producing a main effect of sex (F (1, 82) = 8.0, p < 0.01, η² = 0.09) and a Day × Sex interaction (F (5, 410) = 6.3, p < 0.01, η² = 0.07).

FIGURE 5.

No differences in learning acquisition were observed among females (A). However, males exposed to combined EtOH during the neonatal and adolescent periods exhibited smaller heading angles compared subjects exposed to neonatal EtOH alone (B). A similar pattern was seen in thigmotaxis. There were no differences in thigmotaxis among female subjects (C). However, males exposed to EtOH during both the neonatal and adolescent periods showed significantly less thigmotaxis during visuospatial learning training compared to those exposed to neonatal EtOH exposure alone (D). *neonatal EtOH + adolescent EtOH < neonatal EtOH + sham.

Accuracy and path length to find the platform can be affected by a subject's search strategy. Initially, when subjects are first placed into the tank, they swim around the perimeter close to the wall of the maze, a behavior referred to as thigmotaxis. Thigmotaxis can also indicate anxiety levels, as subjects that are stressed stay close to the maze wall rather than venturing into the center of the water tank. In general, subjects spent less percent time in thigmotaxis during training, producing a main effect of Day (F (5, 410) = 41.2, p < 0.01, η² = 0.34), Trial (F (3, 246) = 91.0, p < 0.01, η² = 0.53), and a Day*Trial interaction (F (15, 1230) = 16.7, p < 0.01, η² = 0.17) (Figure 5C,D). Females spent more time swimming around the perimeter than males and showed less reduction in thigmotaxis over training days, resulting in a main effect of sex (F (1, 82) = 13.4, p < 0.01, η² = 0.14), a Day*Sex interaction (F (5, 410) = 5.1, p < 0.01, η² = 0.06), and a Day*Trial*Sex interaction (F (15, 1230) = 1.9, p < 0.05, η² = 0.02). Thigmotaxic behavior persisted during training in males exposed to only EtOH during the neonatal period, whereas thigmotaxic behavior declined in males exposed to the combination of neonatal and adolescent EtOH. Follow‐up analyses indicated that an interaction of NeoEtOH × AdoEtOH exposure was evident on Days 3–6 of acquisition in males (F (1, 41) = 4.5, p < 0.05, η² = 0.10). Student–Newman Keul's post‐hoc analyses confirmed that this interaction was due to the marked decrease in thigmotaxis among males in the combination exposure group (p < 0.05) in comparison to the thigmotaxic levels observed in males exposed to EtOH during the neonatal period, suggesting that these subjects may be utilizing different search strategies and/or exhibiting different levels of anxiety and risk‐taking. Similar effects were found when data for percent of path length traveled close to the perimeter were analyzed. Notably, this pattern is similar to differences in heading angles, suggesting that their thigmotaxic behaviors may impact heading angles.

The swimming speed (m/s) of all subjects were also analyzed to identify any potential group differences in performance variables (Figure 6). Swimming speeds generally decreased over time, as indicated by significant effects of Day (F (5, 410) = 30.8, p < 0.001, η² = 0.27) and Trial (F (3, 246) = 14.1, p < 0.001, η² = 0.15), as well as an interaction of Day*Time (F (15, 1230) = 2.6, p < 0.01, η² = 0.03; data not shown). In addition, significant interactions of Trial × Sex (F (3, 246) = 4.6, p < 0.01, η² = 0.05), Sex*NeoEtOH (F (1, 82) = 6.3, p < 0.05, η² = 0.07), and Day × Trial × Sex × NeoEtOH (F (15, 1230) = 1.8, p < 0.05, η² = 0.02) were observed. When data were analyzed separately for each sex, there were no significant differences in speed among females. In contrast, males exposed to EtOH during the neonatal period swam slower in the latter half of acquisition (Day × Trial × NeoEtOH: F (15, 645) = 2.1, p < 0.05, η² = 0.05). Importantly, these data suggest that motor function did not impact performance, nor did the pattern relate to the other outcome measures. Moreover, there were no group differences in performance on the visible platform trials.

FIGURE 6.

No differences in swimming speed were observed among females (A). Males (B) exposed to EtOH during the neonatal period swam slower in the latter half of acquisition.*Neonatal EtOH < no neonatal EtOH.

During the probe (test) trial, there were no differences on time spent in target quadrant, time spent in target area, or number of passes through the target, suggesting that neither EtOH exposure period impaired visuospatial memory performance (data not shown).

Trace fear conditioning

During trace fear conditioning, females were less active than males during the baseline (F (1, 82) = 5.8, p < 0.05, η² = 0.07). Freezing increased with learning in both the sexes. However, subjects exposed to neonatal EtOH, alone or in combination with adolescent EtOH exposure, exhibited less freezing over subsequent trials, producing a Trial × NeoEtOH interaction (F (4, 328) = 3.4, p < 0.01, η² = 0.04;Figure 7A). Follow‐up simple effects analyses were conducted for each Trial; overall, groups exposed to neonatal EtOH exhibited slower acquisition, with less freezing during and Trial 4 (F (1, 82) = 3.9, p = 0.05, η² = 0.05).

FIGURE 7.

Neonatal EtOH exposure impaired acquisition of trace fear conditioning, particularly when combined with adolescent EtOH exposure. When collapsed across groups, freezing was significantly reduced among subjects exposed to neonatal EtOH on trials 3 and 4. Neonatal EtOH exposure reduced freezing during trace and ITI events. * = Neonatal EtOH < no neonatal EtOH, p < 0.05. ** = Adolescent EtOH < no Adolescent EtOH, p < 0.05.

Each trial consisted of the CS, CS–US Trace interval and ITI (each referred to as event; see Figure 7B,C). The lowest freezing occurred during each CS and highest freezing during each trace interval, producing effects of Trial (F (4, 328) = 75.4, p < 0.01, η² = 0.48), Event (F (2, 164) = 124.4, p < 0.001, η² = 0.60), and a Trial × Event interaction (F (8, 656) = 66.6, p < 0.001, η² = 0.45). There was also an interaction of Event × NeoEtOH (F (2, 164) = 2.92, p = 0.05, η² = 0.03), as neonatal EtOH exposure reduced freezing during trace and ITI events. Follow‐up analyses of CS trials, trace trials, and ITI trials were conducted separately to better elucidate these interactions. Subject's freezing levels increased over 5 CS administrations, producing a main effect of CS trial (F (4, 328) = 13.1, p < 0.001, η² = 0.14). A CSTrial* Sex × AdoEtOH interaction also approached significance (F (4, 328) = 2.2, p = 0.06, η² = 0.03); follow‐up analyses indicated that female subjects that received EtOH during the adolescent period (alone or in combination) had lower levels of freezing than other subjects during CS trial 4 (F (1, 41) = 5.6, p < 0.05, η² = 0.12).

Similar to CS Trial freezing, subjects also increased freezing level through subsequent trace intervals, producing a main effect of trace Trial (F (4, 328) = 38.1, p < 0.001, η² = 0.32). A Trial × NeoEtOH interaction was also significant (F (4, 328) = 2.4, p = 0.05, η² = 0.03); follow‐up analyses for each trace trial illustrated that subjects given EtOH during the neonatal period (alone or in combination with adolescent EtOH exposure) froze less than their Sham counterpart groups; this was particularly evident during trace trial 3 (F (1, 82) = 5.4, p < 0.05, η² = 0.06).

Subjects also increased freezing levels over subsequent inter‐trial intervals (ITIs), producing a main effect of ITI (F (4, 328) = 87.8, p < 0.001, η² = 0.52). Subjects that received neonatal EtOH exposure had lower freezing during the ITIs, indicated by an interaction of ITI*NeoEtOH (F (4, 328) = 2.6, p < 0.05, η² = 0.03), driven by effects on ITI trials 4 (F (1, 82) = 3.5, p = 0.06, η² = 0.04) and 5 (F (1, 82) = 4.0, p < 0.05, η² = 0.05). In sum, the slower acquisition of a learned fear response following NeoEtOH exposure was most robust during the trace period.

During the context test (Figure 8A,B), subjects showed increasing levels of freezing over the first 4 min, producing a main effect of Time (F (4, 328) = 14.0, p < 0.001, η² = 0.15). Male subjects exhibited higher levels of freezing than female subjects throughout the session (F (1, 82) = 10.4, p < 0.01, η² = 0.12). Among females, the 3‐way interaction of Time × NeoEtOH × AdoEtOH approached significance (F (4, 164) = 2.1, p = 0.08, η² = 0.05). Follow‐up analyses indicated that female offspring exposed to neonatal EtOH alone froze less than those exposed to the combination (F (1, 19) = 5.0, p < 0.05, η² = 0.21). In addition, female offspring exposed to neonatal EtOH alone froze less than Sham+Sham controls during minute 4 (F (1, 19) = 5.14, p < 0.05, η² = 0.21).

FIGURE 8.

Females exposed to EtOH during the neonatal period only froze less during minutes 3 and 4 of the context test (A). No differences were seen among male offspring (B). Females exposed to alcohol during adolescence froze significantly more during the trace period of the CS test than other groups (C). Males froze more than females during the continuous tone, p < 0.05 (D). * = Neonatal EtOH + sham < Neonatal EtOH + Adolescent EtOH, p < 0.05. ** = Neonatal EtOH < sham + sham, p < 0.05. *** = Adolescent EtOH > no adolescent EtOH, p < 0.01.

During the CS test (Figure 8C,D), males froze more on average than female subjects during the baseline (F (1, 82) = 8.5, p < 0.01, η² = 0.09). Subjects' freezing levels remained low during the discrete CS, with no effects of sex or either EtOH exposure. However, during the trace period, a Sex × AdoEtOH interaction was evident (F (1, 82) = 5.6, p < 0.05, η² = 0.06). Follow‐up analyses indicated that females exposed to EtOH during adolescence (alone or in combination) froze significantly more than their counterpart groups during the trace period, producing a main effect of AdoEtOH (F (1, 41) = 8.4, p < 0.01, η² = 0.17); no differences were seen among males. No sex or group differences were observed during the ITI. During the continuous tone presentation, male subjects had higher freezing levels than females, producing a main effect of sex (F (1, 82) = 8.25, p < 0.01, η² = 0.09).

Brain weights

Data were not collected from 5 subjects due to problems with brain tissue collection (NeoEtOH + AdoEtOH: 1 female; 1 NeoEtOH + Sham: 1 female; NeoEtOH + Sham: 1 male; Sham + Sham: 1 female, 1 male). There were no differences in body weights among either EtOH exposure group on the day of brain collection; however, male subjects weighed more than females overall (F (1, 78) = 434.7, p < 0.01, η² = 0.84).

Male subjects had greater brain weights (F (1, 78) = 70.1, p < 0.01, η² = 0.47), forebrain weights (F (1, 78) = 63.6, p < 0.01, η² = 0.45) and cerebellar weights (F (1, 77) = 19.3, p < 0.01, η² = 0.20) compared to females (Table 2). Subjects exposed to EtOH during the neonatal period (alone or in combination with adolescent EtOH exposure) had significantly smaller total brain weights (F (1, 78) = 8.2, p < 0.01, η² = 0.10), as well as smaller forebrain (F (1, 78) = 8.9, p < 0.01, η² = 0.10) and cerebellar regions (F (1, 77) = 8.3, p < 0.01, η² = 0.10).

TABLE 2.

Overall, female offspring had lower body and brain weights compared to males. Separately, neonatal EtOH exposure decreased total brain weights, including the forebrain and cerebellum.

Exposure group	n	Weights means ± SEM (g)
		Body Weights	Total Brain	Forebrain	Cerebellum
Neonatal EtOH + Adolescent EtOH
Females	10	+255 ± 17.9	+*1.85 ± 0.07	+*1.37 ± 0.05	+*0.27 ± 0.03
Males	10	398 ± 53.5	*1.99 ± 0.10	*1.46 ± 0.08	*0.28 ± 0.03
Neonatal EtOH + Sham
Females	9	+259 ± 28.6	+*1.86 ± 0.08	+*1.37 ± 0.06	+*0.27 ± 0.02
Males	10	424 ± 47.7	*2.01 ± 0.07	*1.49 ± 0.06	*0.30 ± 0.01
Sham + Adolescent EtOH
Females	12	+246 ± 13.7	+1.90 ± 0.08	+1.41 ± 0.07	+0.27 ± 0.02
Males	12	415 ± 41.4	2.06 ± 0.07	1.50 ± 0.04	0.31 ± 0.02
Sham + Sham
Females	11	+263 ± 24.9	+1.90 ± 0.11	+1.41 ± 0.03	+0.29 ± 0.02
Males	11	427 ± 39.1	2.06 ± 0.09	1.52 ± 0.05	0.31 ± 0.02

Note: ⁺ = Females < Males, p < 0.01. * = Neonatal EtOH < no Neonatal EtOH, p < 0.01.

DISCUSSION

The present study examined how a history of early alcohol exposure might influence the effects of alcohol during adolescence on behavioral outcome, given that early alcohol exposure may be a risk factor for later alcohol use. Results indicate that the effects of alcohol are sex‐specific and depend on the developmental timing of exposure. Alcohol exposure during the 3rd trimester equivalent reduced gross brain size, increased exploration and time in the center of the open field, and impaired trace fear conditioning acquisition. Females were more robustly affected, exhibiting hyperactivity in the open field center and deficits in contextual memory. Unique effects of alcohol exposure during adolescence were also evident only among females, as they exhibited less locomotor activity and time spent in the center of the open field, and enhanced CS learning on the fear conditioning task. Importantly, the combination of neonatal and adolescent alcohol exposure had the greatest impact on emotional responding, enhancing time spent in the center of the chamber in both sexes and reducing thigmotaxis in the spatial learning task among males (see summary in Table 3).

TABLE 3.

Summary of main findings for each behavioral task.

Task	Measure	Results
Open‐Field Activity	Locomotion	↓ ♀ Adolescent EtOH only
	Rearing	↑ ♀ Neonatal EtOH
	Center Locomotion	↑ ♀ Neonatal EtOH ↓ ♀ Adolescent EtOH only
	Center Entries	↑ Neonatal EtOH
	Center Time	↑ Neonatal EtOH ↑ Combination (less habituation)
Morris Water Maze	Heading Angle	↓ ♂ Combination ↑ ♂ Neonatal EtOH only
	Thigmotaxis	↓ ♂ Combination ↑ ♂ Neonatal EtOH only
Fear Conditioning	Acquisition	↓ Neonatal EtOH
	Context Test	↓ ♀ Neonatal EtOH
	CS Test	↑ ♀ Adolescent EtOH

Open field behaviors can represent locomotion, arousal, exploration and anxiety. Neonatal alcohol exposure led to increased number of entries and time spent in the center of the open field, with females also exhibiting increases in rearing and center locomotor activity. Past literature has been mixed regarding heightened or unaffected locomotor activity, as well as sexual dimorphisms following early alcohol exposure (Bake et al., 2021; Cullen et al., 2013; Marquardt & Brigman, 2016; Patten et al., 2016; Skorput & Yeh, 2016). In contrast, females exposed to alcohol during adolescence exhibited less locomotor activity overall; similar results have been seen in mice (Maldonado‐Devincci, 2021). Thus, these patterns suggest that neonatal alcohol increases activity and reduces anxiety and/or increases risk‐taking behaviors, whereas adolescent alcohol exposure reduces activity and increases anxiety‐related behaviors. Interestingly, subjects that received a combination of neonatal and adolescent alcohol exposure spent significantly more time in the open field center compared to other groups, even though the distance traveled in the center was similar to that of the neonatal alcohol only exposure group. These results indicate that center time was not simply related to increased locomotor activity, but rather reduced habituation and/or increased risk taking in the combination exposure group, as these subjects spent time in the ecologically most vulnerable area of the open field.

In the Morris water maze, no differences in learning acquisition were observed among groups. Similarly, there was no indication of group differences in motor function or motivation, as indicated by swim speed and visible platform testing. However, there was a striking difference in thigmotaxis, as males exposed only to neonatal alcohol exhibited significantly more thigmotaxis compared to males exposed to the combination of neonatal and adolescent alcohol. This suggests that males in the combination group were more comfortable spending time in open space – a result similar to the behavioral pattern of the combination group seen in open field activity (increased center time). It is also possible that differences in search strategy were used, although one might expect that would translate to varying levels of learning performance.

In the fear conditioning, neonatally exposed subjects (the combination and neonatal only group) showed slower learning during training. This effect was particularly robust among the combined exposure group. Importantly, subjects in the neonatal EtOH alone group showed freezing levels comparable to the control group during the CS Test, indicating that they did form an associative memory between the tone and the shock. However, females within the neonatal alcohol alone group exhibited deficits in contextual learning within the fear conditioning task. Previous work has shown contextual memory impairments following high neonatal alcohol doses (5.25 g/kg/day resulting in BACs of 414 ± 13.4 mg/dl), but not at lower doses comparable to the present study (2.75 g/kg/day) (Dokovna et al., 2013; Jablonski & Stanton, 2014; Murawski & Stanton, 2011). Thus, this study indicates that BACs around 200 mg/dl during the 3rd trimester equivalent can sufficiently disrupt this hippocampal‐dependent associative learning.

Somewhat surprisingly, the adolescent alcohol exposure alone group did not show deficits in learning and memory in the Morris water maze or fear conditioning. Binge‐like adolescent alcohol exposure causes damage to the hippocampus and PFC, often resulting in learning and memory deficits (Ehlers et al., 2013; Seemiller & Gould, 2020). Our findings contrast with others who have reported that 2.5 g/kg/day or higher alcohol exposure delivered in a binge‐like manner during adolescence induces long‐lasting deficits in trace fear conditioning (BACs not reported; Ji et al., 2018; Yttri et al., 2004) and working memory (BACs between 200 and 350 mg/dl; Schulteis et al., 2008). However, others have found no deficits in spatial learning and memory with a binge‐like dose of 5.0 g/kg/day (BACs not reported; Vetreno & Crews, 2012), despite increased neuroinflammation and reversal learning impairments on the Barnes maze. Interestingly, in the present study, females exposed to alcohol during adolescence (alone or in combination with neonatal alcohol exposure) froze more during the CS test, suggesting a strong association between the CS and shock. Nevertheless, the Morris water maze and fear conditioning data indicate that alcohol exposure during the adolescent time period alone (average BAC of 200 mg/dl) did not cause impairments in learning and memory performance in the tasks used in this study.

The combination of neonatal and adolescent alcohol exposure produced notable effects on each task. First, the combination exposure led to reduced habituation in the time spent in the center of the open field. In addition, in the Morris water maze, male subjects in the combination group displayed enhanced accuracy in heading angle over consecutive training days when compared to the neonatal exposure only group. Although reduced heading angles might suggest improved spatial memory, this group did not show shorter path lengths or latencies to find the platform during acquisition, nor did they show improved performance on the probe trial, performing similarly to controls on these outcomes. However, these males displayed the lowest thigmotaxis through acquisition, showing an enhanced willingness to enter the center of the Morris water maze. Thus, in both the open field and in the Morris water maze, the increased center time displayed by subjects in the combination group is a striking effect. In addition, the females in the combination group froze less during fear conditioning acquisition, but showed no impairments in contextual learning and improved performance on CS learning during test trials. This lowered freezing may suggest reduced reactivity to aversive stimuli. Collectively, these data suggest that the combination exposure more specifically affects emotional responding by reducing anxiety or reactivity, and/or increasing risk‐taking.

Although the consequences of combination exposure have not been well studied, stress regulation can be altered by either early alcohol exposure (Lam et al., 2019; Weinberg et al., 2008) or adolescent alcohol exposure (Allen et al., 2011; Logrip et al., 2013; Ogilvie & Rivier, 1997). In fact, others have reported that adolescent alcohol exposure in rodents can reduce anxiety‐related behaviors and increase exploratory behavior in the elevated plus maze (Gass et al., 2014) and in an open field task (Ehlers et al., 2013). Relatedly, these behaviors may also represent increased impulsivity or risk‐taking (Crews et al., 2016), possibly related to altered PFC function, which could translate to a variety of maladaptive behaviors. In fact, both clinical and preclinical studies indicate that adolescent alcohol exposure increases risky deviant and delinquent behaviors, and predicts future substance use (Jacobus et al., 2013; Miller et al., 2016; Spear, 2018). The present data suggest that prenatal alcohol exposure may increase vulnerability to such outcomes following adolescent alcohol exposure. Future research should examine mechanisms that may explain the behavioral patterns seen in this study.

Further investigation is needed to identify the neuropathology underlying the behavioral changes. The present study only examined gross brain changes, showing that neonatal alcohol exposure reduced total brain, forebrain, and cerebellum weights (evident in both the neonatal alcohol group and the combination group). Thus, although the alcohol exposure level was selected to not produce severe pathology, blood alcohol levels of 200 mg/dl during the neonatal period can produce gross neuropathology. In contrast, adolescent alcohol exposure alone did not reduce total or regional brain weights. The existing literature has shown that both neonatal alcohol exposure and adolescent alcohol exposure result in lower volume of various brain regions (Crews et al., 2019). Previous research with adolescents has also found that alcohol leads to cell death and reduced neurogenesis, specifically in the hippocampus (Ehlers et al., 2013; Morris et al., 2010; Pascual et al., 2007) and the PFC (De Bellis et al., 2005; Medina et al., 2008). The brain weight data collected in the present study did not examine the hippocampus and PFC (though the behavioral tasks required both brain regions), and thus targeted analyses of neuropathology are needed.

Future studies should more directly measure anxiety‐related behaviors and stress (such as corticosterone measurements) following alcohol exposure after both developmental periods or in conjunction with behavioral tests. For example, testing with the elevated plus maze task could better differentiate anxiety‐related behaviors and exploratory behaviors among exposure groups. It would also be important to track the estrous cycle in females, as hormones can affect performance (Sayin et al., 2014). In addition, one limitation of the present study was that only a single dose was used at each developmental period. Elucidation of how varying alcohol doses influence interactive effects would better identify the clinical risks of dual exposure. Finally, it should be noted that order of behavioral tests was not counterbalanced and although the order was chosen to reduce carry‐over effects, such effects cannot be discounted.

Relating the findings of an animal study to clinical populations is challenging. Nevertheless, animal studies can provide insight about different domains of functioning that are impacted by alcohol. Overall, the present study indicates that alcohol exposure at 200 mg/dl/day during the 3rd trimester equivalent, but not during the adolescent period, produces gross brain pathology and results in hyperactivity and impairments in trace fear conditioning. The hypothesis that neonatal alcohol exposure would also exacerbate the cognitive effects of subsequent adolescent alcohol exposure was not supported with the exposure parameters used in the current study. However, the early alcohol exposure may affect emotional consequences of adolescent alcohol exposure. Since emotional regulation affects many behavioral domains and life function, it is imperative that we fully understand the consequences of drinking alcohol during adolescence among individuals with a history of the prenatal alcohol exposure.

CONFLICT OF INTEREST

None of the authors have any conflict of interest related to this project.

FUNDING INFORMATION

Supported by NIAAA AA012446 (JDT) and T32 AA007456–38 (KRB).

Risbud, R.D. , Breit, K.R. & Thomas, J.D. (2022) Early developmental alcohol exposure alters behavioral outcomes following adolescent re‐exposure in a rat model. Alcoholism: Clinical and Experimental Research, 46, 1993–2009. Available from: 10.1111/acer.14950