Effects of neonatal alcohol dose and exposure window on long delay and trace eyeblink conditioning in juvenile rats

Behavioural Brain Research 236 (2013) 307–318

Contents lists available at SciVerse ScienceDirect

Behavioural Brain Research journal homepage: www.elsevier.com/locate/bbr

Research report

Effects of neonatal alcohol dose and exposure window on long delay and trace eyeblink conditioning in juvenile rats Nathen J. Murawski 2 , Sarah A. Jablonski, Kevin L. Brown 1 , Mark E. Stanton ∗ Department of Psychology, University of Delaware, Newark, DE 19716, USA

h i g h l i g h t s  Both trace and long-delay eyeblink conditioning are impaired in a rat model of FASD.  First study of different doses and developmental periods of alcohol exposure.  Findings relate to brain injury and eyeblink conditioning deficits in human FASD.

a r t i c l e

i n f o

Article history: Received 16 June 2011 Received in revised form 13 August 2012 Accepted 16 August 2012 Available online 28 August 2012 Keywords: Fetal alcohol spectrum disorder Eyeblink conditioning Cerebellum Hippocampus Memory development

a b s t r a c t Classical eyeblink conditioning has been used to assess learning and memory impairments in both humans and animal model studies of fetal alcohol spectrum disorders (FASD). Gestational exposure to alcohol in humans and its equivalent in rats severely impairs various eyeblink conditioning tasks, but less is known about how these effects are influenced by variables, such as the timing and dose of alcohol exposure. In a series of four experiments, we systematically examine how varying the timing and dose of alcohol exposure impact long delay and trace eyeblink conditioning in juvenile rats, tasks that both depend on a brainstem-cerebellar circuit but differ in that trace conditioning additionally recruits the hippocampus and prefrontal cortex. Using a “third-trimester-equivalent” alcohol exposure model, rats were exposed to a high binge dose of alcohol at one of two alcohol doses over postnatal days (PD) 4–9 or PD 7–9, windows of exposure thought to differentially target the cerebellum and hippocampus. Sham-intubated and untreated rats served as controls. As juveniles, rats from each treatment condition were trained in either a long delay or trace eyeblink conditioning task. Alcohol-exposed rats demonstrated general conditioning impairments compared to controls during long delay conditioning, with more robust impairments in rats exposed to the higher alcohol dose (5.25 g/kg/day) than those that received the lower dose (4.66 g/kg/day). Alcohol-exposed rats showed trace conditioning impairments compared to controls only when the high dose of alcohol was administered over PD 4–9 or PD 7–9. These findings indicate significant learning and memory impairments following neonatal alcohol exposure at both PD 4–9 and PD 7–9. The pattern of impairments across delay and trace conditioning suggest that alcohol disrupts processes that are common to both tasks. These findings are consistent with studies of delay and trace eyeblink conditioning in children with FASD. Future studies of the mechanisms underlying these deficits will further our understanding of brain injury and memory impairments resulting from developmental alcohol exposure. © 2012 Published by Elsevier B.V.

1. Introduction

∗ Corresponding author at: Wolf 132, Department of Psychology, University of Delaware, Newark, DE 19716, USA. Tel.: +1 302 831 0575; fax: +1 302 831 3645. E-mail address: [email protected] (M.E. Stanton). 1 Present address: Department of Psychology, University of Iowa, Iowa, IA 52242, USA. 2 Present address: Center for Behavioral Teratology, San Diego State University, San Diego, CA 92120, USA. 0166-4328/$ – see front matter © 2012 Published by Elsevier B.V. http://dx.doi.org/10.1016/j.bbr.2012.08.025

Learning and memory impairments are consistently reported in children and adolescents with known prenatal exposure to alcohol [1–7]. In addition to a general reduction in cognitive function (e.g., IQ) [8], children with fetal alcohol spectrum disorders (FASD) show impairments on behavioral tasks that reflect specific domains of learning and memory. For example, children with FASD show significant acquisition deficits during classical eyeblink conditioning [3,4] that likely result from cerebellar abnormalities commonly demonstrated in this subject population [9,10]. Additionally, children with FASD show difficulty on spatial learning tasks [5–7],

308

N.J. Murawski et al. / Behavioural Brain Research 236 (2013) 307–318

which may reflect abnormal hippocampal development following gestational alcohol exposure [7,11,12]. These behavioral and anatomical outcomes are also reported in animal models of FASD [13–16], which provide experimental control of the significant variables contributing to alcohol-related teratogenicity (e.g., the developmental timing and dose of alcohol exposure) [15–18]. The use of behavioral tasks like eyeblink conditioning in animal models of FASD may further our understanding of how the timing and dose of alcohol exposure contribute to domain-specific behavioral impairments. Classical eyeblink conditioning offers a potential mechanism by which to examine both cerebellar and hippocampal function in the alcohol-exposed brain. During delay conditioning, a conditioning stimulus (CS; e.g., tone) overlaps and co-terminates with an unconditioned stimulus (US; e.g., airpuff to the eye or periocular shock). After repeated pairings, the previously neutral CS comes to elicit a conditioned response (CR; e.g., eyelid closure) prior to, or in the absence of, the US. The neural circuitry underlying delay eyeblink conditioning involves a brainstem-cerebellar pathway that is conserved across a number of species, including rodents and humans [19–24]. By imposing a stimulus-free “trace” interval between the CS and US, forebrain structures, such as the hippocampus, in addition to the underlying brainstem-cerebellar circuit, are needed for acquisition [25,26]. An additional distinction between delay and trace conditioning concerns the relative timing between CS and US onsets (the interstimulus interval [ISI]), which is usually longer during trace conditioning. By matching the ISI in delay (i.e., long delay) with that of trace conditioning, the behavioral ontogeny and acquisition rates are similar across tasks [27,28]; importantly, hippocampal lesions impair trace conditioning more than longdelay conditioning [25]. Long delay and trace eyeblink conditioning, therefore, provide an opportunity to examine both cerebellar and hippocampal function, respectively, in animal models of FASD. The cerebellum and hippocampus are vulnerable to the neurotoxic effects of alcohol, especially when alcohol exposure is limited to the third-trimester equivalent in the rat (i.e., the first postnatal week) [15,16,29]. Purkinje cells within the cerebellum appear most vulnerable to the effects of alcohol prior to the 7th postnatal day (PD) [15,30], while alcohol exposure limited to either PD 4–9 or PD 7–9 results in CA1 pyramidal cell reduction within the hippocampus [14,16]. Limiting alcohol exposure to PD 7–9, therefore, is likely to result in significant hippocampal targeting relative to the cerebellum and may result in differential performance on long delay and trace eyeblink conditioning. Recent studies in humans with FASD and in rodent models have utilized delay and trace eyeblink conditioning paradigms, demonstrating comparable conditioning impairments regardless of task [3,31]. Here we further explore the impact of developmental alcohol exposure on trace and long delay eyeblink conditioning in juvenile rats neonatally exposed to alcohol by manipulating the timing (PD 4–9 or PD 7–9) and dose (5.25 or 4.66 g/kg/day) of alcohol exposure. We hypothesized that rats exposed to alcohol over PD 4–9 would show impairments on both trace and long delay tasks, while those exposed over PD 7–9 would show greater impairments during trace than during long delay conditioning. We also hypothesized that greater deficits would appear in rats that received higher doses of alcohol.

2. Experiment 1A: the effects of a high binge dose of alcohol (5.25 g) over PD 4–9 on long delay and trace eyeblink conditioning Neonatal alcohol exposure (PD 4–9) leads to cell loss in both the cerebellum and the hippocampus [15,16]. Here we ask what effects exposure to a high dose of alcohol (5.25 g/kg/day) over the neonatal

period (PD 4–9) has on two task variants of eyeblink conditioning (EBC; trace and long delay) that both rely on a known cerebellar circuit but differ in their sensitivity to hippocampal disruption [19,25]. 2.1. Materials and methods 2.1.1. Subjects The subjects were 90 Long Evans rats (46 males and 44 females) from 29 litters bred at the University of Delaware. Breeder rats were housed overnight and if an ejaculatory plug was found the following morning that day was designated as gestational day (GD) 0. Pregnant females were housed in clear polypropylene cages (45 cm × 24 cm × 21 cm) with standard bedding and ad lib. access to water and rat chow. The date of birth was determined by checking for births during the light cycle (12:12) and, if newborn pups were found, that day was designated as postnatal day (PD) 0; all births occurred on GD 22. On PD 2, litters were transported from the breeding facility to be housed in the animal housing facility. Litters were culled to 8 pups (4 males and 4 females when possible) and received subcutaneous injections of a nontoxic black ink into one or more paws to aid in identification on PD 3. On PD 21, pups were weaned and housed with same-sex litter mates in 45 cm × 24 cm × 17 cm cages with ad lib. access to water and rat chow. On PD 24, rats were individually housed in small white polypropylene cages (24 cm × 18 cm × 13 cm) for the remainder of the study. 2.1.2. Alcohol dosing Pups were exposed to a 5.25 g/kg/day single binge dose of alcohol (5.25 g), sham intubated (SI), or left undisturbed (UD) over PD 4–9 following procedures previously reported [32,33]. Within a given litter, pups were randomly assigned to receive alcohol or sham intubations. UD rats came from separate litters, which were left undisturbed with the exception that body weights were obtained on PD 4 and PD 9. Starting on PD 4, pups were briefly separated from their mothers and placed into Lexan containers placed over a heating pad (GE model #E12107) that was turned to the lowest setting to maintain body temperatures during separation. Alcohol was mixed with a custom milk formula [34] at a dose of or 5.25 g/kg/day (23.93% v/v). Pups were weighed each morning to determine the volume of solution to be administered at 0.0278 mL/g. Alcohol administration via intragastric intubation involved passing PE10 tubing lubricated with corn oil down the esophagus and into the stomach. The tubing was connected to a 1 mL syringe that was used to infuse the solution over a 15 s interval. Alcohol was administered in a single binge dose, followed 2 h later by an equal volume of milk-only solution. A second milk-only infusion was administered only on PD 4. Sham intubations involved an identical process with the exception that no solution was infused. 2.1.3. Blood alcohol concentration analysis Two hours after alcohol/sham administration on PD 4, a 20 ␮L blood sample was collected from a small tail clip using a heparinized capillary tube. Blood samples from SI pups were disposed, while those collected from alcohol-exposed pups were centrifuged and plasma was collected and stored at −20 ◦ C. Blood alcohol concentrations (BACs) were determined using an Analox GL5 Analyzer (Analox Instruments, Luneburg, MA) as previously described [17]. Briefly, the rate of oxidation of alcohol in each plasma sample was measured. BACs (expressed in mg/dl) were calculated based on comparisons to known values of an alcohol standard solution. 2.1.4. Eyeblink surgery Subjects received eyeblink surgeries on PD 24 as previously described [17,35,36]. Pups were anesthetized with a ketamine/xylazine cocktail (87 mg/kg ketamine/13 mg/kg xylazine) with an intraperitoneal injection at a volume of 0.8 mL/kg. The scalp was cleaned with betadine and 70% isopropyl alcohol prior to incision and head stage implantation. The head stage consisted of differential electromyography (EMG) electrodes threaded through the upper left eyelid muscle, a bipolar stimulating electrode placed subdermally caudal to the left eye (to deliver the US), and a ground wire curled up and placed subdermally towards the back of the incision. The head stage was secured to the skull with dental acrylic that adhered to two 15 mm strips of galvanized steel wires implanted into the skull to act as anchors (placed in front and behind the head stage). Following surgery, an antiseptic opthalmic ointment was placed onto the eyes. The pups were then placed onto a heating pad set to the lowest temperature as they recovered from the anesthetic (∼30–45 min). 2.1.5. Apparatus Eyeblink conditioning occurred in one of 16 sound-attenuated conditioning chambers (BRS/LVE, Laurel, MD) as previously described [35,37]. Rats were placed into a stainless wire mesh cage within the conditioning chamber during testing. A fan provided background noise (70 dB). The auditory CS was produced by one of two speakers placed within each conditioning chamber. The current study used an 80 dB, 2.8 kHz tone presented for either 380 or 980 ms. The US was produced by a constantcurrent, 60 Hz square wave stimulator (World Precision Instruments, Sarasota, FL) set to deliver a 1.5 mA, 100 ms periocular shock. During conditioning sessions a wire lead that passed through the opening of the conditioning chambers was secured to

N.J. Murawski et al. / Behavioural Brain Research 236 (2013) 307–318

309

activity that crossed this baseline prior to CS onset were examined for signal-tonoise integrity and, if deemed noisy, were excluded from further analysis. If more than 30% of trials for any given animal were determined to be noisy, that animal was excluded from further analysis. CR percentage and amplitude measures are reported from paired CS-US trials. In the absence of a CR the CR amplitude was measured as “zero amplitude” [36]. 2.1.8. Statistical analysis Data were statistically analyzed using Statistica 8 software using analyses of variance (ANOVAs) and post hoc tests (Newman-Keuls) unless otherwise noted. Significance was set to p ≤ 0.05. The two eyeblink conditioning tasks (D880 and T500) were analyzed separately. ANOVA involved the between-subjects factors of treatment (5.25 g, SI, UD) and the within-subjects factor of sessions. Behavioral data were collapsed across sex because this variable did not alter the effect of alcohol treatment in this study. Although adult female rats (>PD 60) show enhanced trace eyeblink conditioning acquisition compared to adult males, these sex differences are not evident in pubescent (PD 35–40) or prepubescent (PD 25–29) rats [38]. Body weights recorded on PD 4, PD 9, and PD 26 were analyzed between sex and treatment.

Fig. 1. Schematic depiction of long delay (D880; bottom) and trace (T500; top) eyeblink conditioning paradigms with matching interstimulus intervals (ISI). Solid lines represent the onset and deration of the tone conditional stimulus (CS) and the shock unconditional stimulus (US). During D880, the tone CS occurs over 980 ms, overlapping and co-terminating with a 100 ms 1.5 mA shock US. During T500, the tone CS occurs over 380 ms and is separated from the onset of the shock US by a 500 ms stimulus-free “trace” interval. Dotted lines show that D880 and T500 share an 880 ms ISI.

rat’s head stage. The wire lead was connected to a commutator that allowed the rats to freely move about the chamber during the conditioning sessions. 2.1.6. Design and procedures On PD 25 (the day following eyeblink surgery) subjects were transported to the testing rooms, weighed, and briefly connected to the conditioning equipment for a “puff test” that involved the experimenter blowing three quick puffs of air towards the left eye. This allowed for an assessment of the EMG signal quality independent of the conditioning procedures. Subjects from each of the three treatments (5.25 g, SI, and UD) were assigned to one of two behavioral conditions (T500 or D880, see below) such that no more than one pup per sex per treatment per litter was assigned to a single behavioral condition. Starting on PD 26, subjects received six eyeblink conditioning sessions based on previously published procedures [25,28]. Sessions consisted of 100 trials per session in 10 blocks of 10 trials. The intertrial interval averaged 30 s, ranging from 18 to 42 s in a pseudorandom fashion. For each block of 10 trials, 9 involved pairing an 80 dB tone CS with a 100 ms 1.5 mA periocular shock US. Every 10th trial was a non-reinforced CS-alone trial. Subjects received two conditioning sessions per day separated by 5 h ± 30 min, typically between 9–11 am and 2–4 pm daily. Prior to the first conditioning session, subjects were weighed and placed into the conditioning chambers for a 10 min acclimation period. For the subjects that received trace conditioning (T500) a 500 ms “trace interval” separated a 380 ms tone CS from the 100 ms periocular shock US. For subjects receiving long delay conditioning (D880) a 980 ms tone CS overlapped and co-terminated with a 100 ms periocular shock US. The interstimulus interval (ISI; the time between CS onset and US onset) was 880 ms for both T500 and D880 groups. Fig. 1 provides a schematic comparing the two conditioning procedures. 2.1.7. Data analysis As described previously [25,28], data collected during conditioning sessions included a 1400 ms trial epic acquired from rectified EMG signals sampled in 3.5 ms bins. On paired trials, each trial epic was further divided into 5 component parts: (1) a pre-CS baseline period of 280 ms; (2) a startle-response period of 80 ms following CS onset; (3) a conditioned response (CR) period of 800 ms; (4) an “adaptive CR” period (the 200 ms prior to US onset); and (5) the unconditioned response (UR) period lasting 140 ms following US offset (data were not collected during the 100 ms US to avoid stimulus artifacts). On CS-alone test trials, the trial epoch was as described above except that the CR period extended to the end of the trial epoch. The dependent measures analyzed reflect EMG activity occurring during different components of the trial epic that meet specified criteria outlined below. The dependent measures include startle responses (SRs) and SR maximum amplitudes (SMAs), total conditioned responses (%CRs), adaptive conditioned responses (adaptive %CRs; final 200 ms of CR period), conditioned response maximum amplitude for both %CRs and adaptive %CRs (CMAs and adaptive CMAs, respectively), and unconditioned responses (URs) and UR maximum amplitudes (UMAs). All figures represent data taken from adaptive measures, as this measure is sensitive to hippocampal insult [25]. The average EMG activity recorded during the pre-CS period established a baseline amplitude for each trial. An EMG response was defined as exceeding the baseline amplitude by 40 arbitrary units [22,37]. Trials with high baselines or EMG

2.2. Results A total of 16 rats (9 males and 7 females) were excluded from analysis. Eight rats were excluded due to excessively noisy EMG signals (UD = 2; SI = 4; 5.25 g = 2) while an additional seven rats were excluded for having EMG signals that were too low (URs ≤ 100 arbitrary units) to record reliably (UD = 2; SI = 2; 5.25 g = 3). One rat from group 5.25 g (T500) was excluded from analysis for meeting the criteria as a statistical outlier. Outliers were defined a priori as subjects with mean scores on measures of both adaptive %CR and CMA that exceeded ±2 standard deviations from the group mean. The body weight of one rat (UD) was not available for analysis. The behavioral analyses were conducted on the remaining 74 subjects, including 38 rats trained under long delay (UD = 13; SI = 14; 5.25 g = 11) and 36 rats trained under trace EBC (UD = 11; SI = 15; 5.25 g = 10). 2.2.1. Body weight and BACs The body weights and BACs from Experiment 1A are listed in Table 1. All groups gained a significant amount of weight between PD 4 and PD 9. Alcohol exposure produced mild growth retardation during the intubation period. A repeated measures ANOVA with sex and treatment as between-subjects factors and days as a within-subject factor revealed a main effect of treatment [F(2, 67) = 20.45, p < 0.01] and a significant days × treatment interaction [F(2, 67) = 141.72, p < 0.01]. Newman-Keuls posthoc analyses revealed no differences between treatments at PD 4, with significant differences at PD 9; Group 5.25 g pups weighed significantly less than Groups SI and UD, with Group SI weighing significantly more than Group UD. PD 26 body weights analyzed with a factorial ANOVA showed a main effect of sex [F(1, 68) = 7.12, p < 0.01] and a main effect of treatment [F(2, 68) = 8.00, p < 0.01], without an interaction. As can be seen in Table 1, females weighed less than males (F = 60.56 g vs. M = 64.46 g) and Group 5.25 g weighed significantly less than control groups, which did not differ from one another (Group 5.25 g = 57.52 g, Group SI = 65.25 g, Group UD = 63.58 g). BAC analysis was performed on 15 of 21 samples (6 samples were lost due to technical error). The average BACs for Group 5.25 g on PD 4 was 395.80 ± 8.35 mg/dl (Table 1). 2.2.2. Performance measures – SR and UR amplitudes Mean (±SE) amplitudes of unconditioned responses (UMAs; across the first 10 trials of Session 1) and startle responses (SMAs; across Session 1) can be found in Table 2. No significant effects of treatment were observed for UMAs on either task (Fs < 1.8). Analysis of SMAs during long delay conditioning revealed a significant treatment effect [F(2,35) = 5.57, p < 0.01], with Group SI startle amplitudes (33.6 ± 8.8) significantly higher than either Group UD (11.4 ± 3.0) or 5.25 g (7.0 ± 3.3), which did not differ from one another (p > 0.6). SMAs during trace conditioning did not differ

310

N.J. Murawski et al. / Behavioural Brain Research 236 (2013) 307–318

Table 1 Mean (±SEM) body weights (g) obtained on the first and last day alcohol was administered and at beginning of eyeblink conditioning (session 1) in undisturbed (UD), sham-intubated (SI), or rats exposed to alcohol at either 5.25 g or 4.66 g (i.e., g/kg/day) over postnatal days (PD) 4–9 or 7–9. Blood alcohol contents (BACs; mg/dl) were obtained from alcohol-exposed pups 2 h following their first alcohol dose on either PD 4 or PD 7. Experiment

Body weight

BACs (mg/dl) Males

Females

Treatment UD SI 5.25 g

PD 4 9.98 ± 0.3 10.32 ± 0.2 10.72 ± 0.2

PD 9 17.90 ± 0.4# 19.71 ± 0.4* 14.10 ± 0.4* , #

PD 26 65.8 ± 1.8 67.5 ± 2.2 59.8 ± 2.2

PD 26 61.30 ± 1.9 63.40 ± 1.8 54.40 ± 2.8* , #

BACs N/A N/A 395.8 ± 32.3

Exp 1B

UD SI 4.66 g

10.78 ± 0.2 10.75 ± 0.1 10.85 ± 0.1

19.44 ± 0.5 19.65 ± 0.4 16.76 ± 0.3* , #

71.09 ± 2.4 71.92 ± 1.7 68.19 ± 1.0

70.87 ± 1.9 66.38 ± 1.4 66.89 ± 2.2

N/A N/A 325.7 ± 15.7

Exp 2A

UD SI 5.25 g

15.34 ± 0.6 15.17 ± 0.6 14.92 ± 0.5

18.74 ± 0.6 20.21 ± 0.5 15.82 ± 0.7* , #

63.0 ± 7.4 74.0 ± 2.3 65.6 ± 3.8

67.93 ± 2.3 67.27 ± 1.6 57.43 ± 3.1

N/A N/A 416.2 ± 9.9

Exp 2B

UD SI 4.66 g

15.31 ± 0.6 15.68 ± 0.6 16.43 ± 0.5

18.51 ± 0.7 20.32 ± 0.7 18.70 ± 0.6#

66.80 ± 2.2 74.30 ± 2.5 70.64 ± 3.3

67.20 ± 2.1 64.83 ± 2.8 64.83 ± 2.9

N/A N/A 305.10 ± 22.6

Exp 1A

* #

Denotes significant difference from Group UD. Denotes significant difference from Group SI.

among treatment groups (F < 1.4). SMAs and UMAs are a measure of CS and US processing, respectively. The significantly higher SMAs in Group SI during long delay conditioning compared to the other groups contrasts with a lack of group differences during trace conditioning, from the other experiments outlined below (Exps. 1B, 2A, and 2B), as well as the results of numerous studies examining neonatal alcohol and eyeblink conditioning [17,35,36]. The absence of group differences, at least between Groups UD and 5.25 g, on these measures rule out basic performance deficits in alcohol-exposed rats and suggest that any group differences in the measures reported below (i.e., adaptive CR% and adaptive CMA) are likely to be specific to associative learning processes. 2.2.3. Long delay eyeblink conditioning Rats from all treatments showed evidence of conditioning over the 6 sessions, with increased adaptive %CRs and adaptive CMAs during Session 6 relative to Session 1 (ps < 0.01). Neonatal alcohol exposure produced acquisition deficits during long delay conditioning (Fig. 2, upper left panel). Analysis of adaptive %CR revealed a significant main effect of treatment [F(2,35) = 8.35, p < 0.01] and a significant sessions × treatment interaction [F(10,175) = 2.88, p < 0.01]. Newman Keuls post hoc

analysis of the sessions × treatment interaction revealed significant reductions in adaptive %CR in Group 5.25 g during the first four sessions compared to Group UD (ps < 0.03) and during the first two sessions compared to Group SI (ps < 0.01); Groups UD and SI did not differ from one another (ps > 0.4). Group 5.25 g also showed significantly reduced adaptive CMAs compared to Groups UD and SI (Fig. 2, bottom left panel). An analysis of adaptive CMAs revealed a significant main effect of treatment [F(2,35) = 7.38, p < 0.01], with significantly lower adaptive CMAs in Group 5.25 g compared to Groups UD and SI (ps < 0.03), which did not differ from one another (p > 0.13). Analysis also revealed a significant sessions × treatment interaction [F(10,175) = 3.19, p < 0.01]. Group 5.25 g showed significant reductions in adaptive CMAs compared to Group UD over the last four sessions of conditioning (ps < 0.03). Group differences between 5.25 g and SI did not reach significance (ps > 0.13). During Session 6, Group UD showed significantly higher adaptive CMAs compared to both Groups SI (p < 0.04) and 5.25 g (p < 0.01). 2.2.4. Trace (T500) eyeblink conditioning Rats from each treatment group showed evidence of trace eyeblink conditioning acquisition, with increased adaptive %CRs and CMAs during Session 6 compared to Session 1 (ps < 0.01). Analysis

Table 2 Mean (±SEM) eyeblink conditioning performance measures include 1st block (first 10 trials) unconditioned response maximum amplitudes (UMAs) and startle response maximum amplitudes (SMAs) from the first session in undisturbed (UD), sham-intubated (SI), or rats exposed to alcohol at either 5.25 g or 4.66 g (i.e., g/kg/day) over postnatal days (PD) 4–9 or 7–9. D880 = Long Delay; T500 = Trace. Experiment

Animal/Group

Performance measure

D880

T500

D880

Treatment UD SI 5.25 g

n= 13 14 11

n= 11 15 10

UMA 729.1 ± 65.7 800.3 ± 55.1 673.2 ± 73.9

SMA 11.4 ± 3.0# 33.6 ± 8.8* 7.0 ± 3.3#

Exp 1B

UD SI 4.66 g

13 14 11

11 15 10

768.2 ± 45.6 671.3 ± 74.7 697.2 ± 93.1

41.0 ± 8.2 25.5 ± 6.6 13.0 ± 4.4*

687.8 ± 79.6 817.9 ± 50.1 704.1 ± 79.4

25.1 ± 12.7 54.5 ± 16.5 14.3 ± 3.9

Exp 2A

UD SI 5.25 g

8 12 8

8 10 9

758.6 ± 59.3 623.6 ± 73.1 621.6 ± 81

137.62 ± 60.1 132.1 ± 50.8 30.9 ± 12.6

707.7 ± 87.6 641.7 ± 46.4 675.4 ± 95.8

61.3 ± 30.8 155.3 ± 56.6 74.6 ± 36.1

Exp 2B

UD SI 4.66 g

14 13 12

9 12 14

792.0 ± 59.1 658.6 ± 66.3 653.5 ± 67.2

37.6 ± 13.4 20.6 ± 6.1 18.9 ± 6.9

572.0 ± 69.1 683.9 ± 59.2 653.4 ± 65.4

25.8 ± 7.7 41.3 ± 15.3 9.6 ± 3.0

Exp 1A

* #

Denotes significant difference from Group UD. Denotes significant difference from Group SI.

T500 UMA 672.6 ± 59.8 825.29 ± 52.8 782.6 ± ± 68.3

SMA 53.4 ± 24.8 32.9 ± 11.1 14.9 ± 3.7

N.J. Murawski et al. / Behavioural Brain Research 236 (2013) 307–318

311

Fig. 2. Mean (±SEM) percentage of adaptive CR percentage (upper panel) and CR peak amplitude (lower panel) during long delay (left) and trace (right) eyeblink conditioning in undisturbed (UD), sham-intubated (SI), or alcohol-exposed (5.25 g) juvenile pups. Alcohol or sham intubations occurred on postnatal days (PD) 4–9 and 6 sessions of either long delay or trace eyeblink conditioning occurred on PD 26–28.

of adaptive %CRs demonstrated a significant sessions x treatment interaction [F(10,165) = 2.09, p < 0.03] without a main effect of treatment (F < 0.7; Fig. 2, upper right panel). Newman Keuls post hoc analysis of the sessions x treatment interaction demonstrated that Groups UD and SI never differed in any session. Group 5.25 g differed from Groups SI and UD in Session 1 (p < .01 against both) and Session 2 (p < .05 vs. SI; p < .01 vs UD) but not thereafter. Across sessions, all groups improved between Sessions 2 and 3 but only the Group 5.25 g improved from Session 3 to 4 (an indication of delayed acquisition). There were no treatment or interaction effects in the adaptive CMA measure (Fs < 1.5).

consistent with a general tendency for ceiling effects to appear late in training in the %CR measure and “floor effects” to appear early in training in the amplitude measure during eyeblink conditioning at these long ISIs [25]. Alcohol-induced acquisition impairments were more robust during long delay than trace conditioning. Importantly, these deficits result from conditioning impairments that cannot be explained by altered CS or US processing in alcoholexposed rats.

2.3. Summary of experiment 1A

3. Experiment 1B: the effects of a moderately high binge dose of alcohol (4.66 g) over PD4-9 on long delay and trace eyeblink conditioning

Juvenile rats exposed to a 5.25 g/kg/day of alcohol over PD 4–9 show significant acquisition deficits during long delay eyeblink conditioning, with some impairment evident during trace conditioning. Specifically, alcohol-exposed rats showed reductions on adaptive %CR measures early during acquisition (during long delay and trace EBC) while reductions on adaptive CMA measures were most apparent during later training sessions when compared to controls (during long delay EBC). This difference across measures is

Rats exposed to alcohol doses less than 5.00 g/kg/day over the neonatal period show deficits on complex eyeblink conditioning tasks, such as ISI discrimination and discrimination reversal learning [17,35,36], although performance on single-cue eyeblink conditioning tasks remains unimpaired [39]. Experiment 1B examined the consequences of a lower dose of alcohol (4.66 g/kg/day) over the neonatal period (PD 4–9) on trace and long delay eyeblink conditioning.

312

N.J. Murawski et al. / Behavioural Brain Research 236 (2013) 307–318

3.1. Materials and methods 3.1.1. Subjects A total of 85 Long Evans rats (42 males and 43 females) from 22 litters were run in Experiment 1B. The housing, animal care and maintenance were the same as in Experiment 1A. 3.1.2. Design and procedures All alcohol dosing, surgery, design, experimental procedures, and data analysis were identical to Experiment 1A, except that a single binge alcohol dose of 4.66 g/kg/day (21.23% v/v) was administered to alcohol-exposed pups over PD 4–9 (4.66 g).

and significant sessions × treatment interaction [F(10,180) = 2.21, p < 0.02], with Group 4.66 g showing reduced adaptive CMAs compared to Group UD over Sessions 4 through 6 (ps < 0.05) and Group SI over Sessions 5 and 6 (ps < 0.03; Fig. 3, bottom left panel). Control groups did not differ from one another across sessions (ps > 0.3). 3.2.4. Trace conditioning Rats from all treatment groups showed significant trace conditioning acquisition across training sessions with greater adaptive %CRs and CMAs in Session 6 compared to Session 1 (ps < 0.01). Rats from the three treatment groups did not differ on measures of adaptive %CRs or CMAs (Fs < 0.9), nor was there a sessions × treatment interaction on these measures (Fs < 1.0; Fig. 3 left panels) (Fig. 3).

3.2. Results 3.3. Summary of experiment 1B A total of 11 animals (5 males and 6 females) were excluded from behavioral analyses. Ten animals were excluded for excessive noise in their EMG signal (UD = 4; SI = 1; 4.66 g = 5) while one UD male rat trained on D880 was excluded for meeting the criteria of a statistical outlier. Further analyses for Experiment 1B were conducted on the remaining 76 rats, including 39 rats run in long delay (UD = 15; SI = 12; 4.66 g = 12) and 37 rats run in trace EBC (UD = 11; SI = 13; 4.66 g = 13). 3.2.1. Body weights and BACs Table 1 summarizes the PD 4, PD 9, and PD 26 body weights. Rats from all treatments gained a significant amount of weight over PD 4 to PD 9 (p < .01). Similar to Experiment 1A, rats exposed to alcohol over PD 4–9 showed reduced body weights on PD 9 compared to the two control groups, whose weights did not differ from one another, as reflected by a main effect of treatment [F(2, 70) = 7.55, p < 0.01] and days × treatment interaction [F(2, 70) = 33.26, p < 0.01]). Body weights on PD 26 did not differ between treatments or by sex (ps > 0.10), suggesting that the initial weight reduction in alcoholexposed rats was temporary. BACs from two alcohol-exposed animals were lost due to technical error. BACs are reported from the remaining 23 rats in Group 4.66 g. The average BACs for Group 4.66 g on PD 4 was 325.72 mg/dl (Table 1). 3.2.2. Performance measures – SR and UR amplitudes Mean (±SE) UR and startle amplitudes (SMA) can be found in Table 2. No significant effects of treatment were observed for UR amplitude on either task (Fs < 1.1). Analysis of SMAs during long delay conditioning revealed a significant treatment effect [F(2,36) = 4.23, p < 0.03], with Group 4.66 g startle amplitude (13.04 ± 4.4) significantly lower than Group UD (41.0 ± 8.2); Group SI did not differ from either group (25.5 ± 6.6; ps > 0.1). Differences in SMAs among treatment groups during trace conditioning did not reach significance (p > 0.06), although Group SI showed higher startle amplitudes (54.5 ± 16.5) than Groups UD (25.1 ± 12.7) and 4.66 g (14.3 ± 3.9); the pattern of CR effects (below) in Group SI do not match their increased startle amplitudes during trace EBC. 3.2.3. Long delay conditioning Rats exposed to a moderately high dose of alcohol over PD 4–9 showed reduced CR acquisition compared to controls during long delay eyeblink conditioning (Fig. 3, left panels). nalysis of adaptive %CRs revealed a significant main effect of treatment [F(2,36) = 5.37, p < 0.01]; the sessions × treatment interaction only approached significance (p > 0.07). Group 4.66 g showed significantly fewer CR responses than both Groups UD and SI during Sessions 1 and 2 (ps < 0.03). Groups UD and SI never differed from one another across all sessions (ps > 0.7). Group 4.66 g also demonstrated significantly reduced adaptive CMAs relative to controls. This was supported by a significant main effect of treatment [F(2,36) = 6.50, p < 0.01]

Experiment 1B examined long delay and trace eyeblink conditioning in juvenile rats exposed to a moderately high binge dose of alcohol (4.66 g/kg/day) over PD 4–9. Rats exposed to alcohol showed significant acquisition impairments during long delay conditioning compared to controls, with reduced adaptive CR responses early in acquisition and reduced adaptive CR amplitude during later sessions. In contrast, acquisition of trace conditioning did not differ among treatment groups. The findings from both Experiment 1A and 1B support our initial hypothesis, that alcohol exposure over PD 4-9, a period when both the cerebellum and hippocampus are vulnerable to the effects of alcohol [15,16], would disrupt long delay and trace eyeblink conditioning in a dose- and task-dependent manner. We did not predict that alcohol effects on delay conditioning would be larger and more consistent than effects on trace conditioning, but this may reflect a predominant role of cerebellar targeting during the PD 4–9 exposure window (see Section 7). The long delay eyeblink conditioning deficits were present at alcohol doses that produced BACs that averaged 395.80 (Experiment 1A) to 325.74 mg/dl (Experiment 1B), with only the higher BACs reached in Exp. 1A affecting trace conditioning acquisition. In the following series of experiments we build upon these initial findings and examine the effects on long delay and trace eyeblink conditioning of alcohol exposure over a latter window of the neonatal period that preferentially targets the hippocampus. 4. Experiment 2A: the effects of a high binge dose of alcohol (5.25 g) over PD 7–9 on long delay and trace eyeblink conditioning Delay and trace eyeblink conditioning share common CS and US pathways but differ on their recruitment of certain brain structures for robust conditioning, such as the cerebellar cortex (during delay conditioning [20,40,41]) and the hippocampus (during trace conditioning [25,42,43]). Like short delay, long delay conditioning is less vulnerable to hippocampal insult than is trace conditioning [25]. Neonatal alcohol exposure in the rat results in significant reductions in Purkinje cells in the cerebellar cortex [15,30,44] and CA1 pyramidal cells in the hippocampus [16,45,46]. Purkinje cells are most vulnerable to the toxic effects of alcohol prior to PD 7 [15,30], while CA1 pyramidal cells remain vulnerable over PD 7–9 [14]. By limiting alcohol exposure to this later neonatal window (i.e., PD 7–9), there should be relative sparing of Purkinje cells in the cerebellar cortex while still causing significant reductions in CA1 pyramidal cells in the hippocampus. In Experiment 2A, we examined long delay and trace conditioning in juvenile rats exposed to a high binge dose of alcohol (5.25 g/kg/day) over PD 7–9. We hypothesized that alcohol-exposed pups would show greater eyeblink conditioning deficits during trace conditioning than during long delay conditioning.

N.J. Murawski et al. / Behavioural Brain Research 236 (2013) 307–318

313

Fig. 3. Mean (±SEM) percentage of adaptive CR percentage (upper panel) and CR peak amplitude (lower panel) during long delay (left) and trace (right) eyeblink conditioning in undisturbed (UD), sham-intubated (SI), or alcohol-exposed (4.66 g) juvenile pups. Alcohol or sham intubations occurred on postnatal days (PD) 4–9 and 6 sessions of either long delay or trace eyeblink conditioning occurred on PD 26–28.

4.1. Materials and methods 4.1.1. Subjects The subjects were 71 Long Evans rats (35 males and 36 females) derived from 13 litters. 4.1.2. Design and procedures All alcohol dosing, surgery, design, experimental procedures and data analysis were identical to Experiment 1A, except that alcohol was only administered from PD 7–9. Accordingly, UD pups were weighed on PD 7 and 9 and two hours after alcohol/sham administration on PD 7, blood samples were taken for BAC analysis. 4.2. Results A total of 12 animals (4 males and 8 females) were eliminated from analysis for having excessive noise in their EMG signals. Four additional animals were excluded from the following analyses for meeting the criteria as a statistical outlier. These included two females from Group UD (1 from T500 and 1 from D880), one female from Group SI (D880), and one male from Group SI (T500). The

analyses were conducted on the remaining 55 subjects, including 28 rats run in long delay (UD = 8; SI = 12; 5.25 g = 8) and 27 run on trace EBC (UD = 8; SI = 10; 5.25 g = 9). 4.2.1. Body weights and BACs All pups gained a significant amount of body weight over the dosing period (PD 7-PD 9; ps < 0.01), although Group 5.25 g experienced moderate growth retardation during the intubation period (Table 1). These effects were supported by a significant main effect of days [F(2,42) = 117.75, p < 0.01] and treatment [F(2,42) = 7.33, p < 0.01] and a significant interaction of treatment × days [F(2,42) = 41.32, p < 0.01]. Newman-Keuls post hoc test of these effects indicated no group difference at PD 7 (p > 0.5), and lower body weights of Group 5.25 g at PD 9 relative to controls (p < 0.001), which did not differ from each other (p > 0.2). Body weights on PD 26 did not differ between treatment groups or by sex (ps > 0.05). BACs were available for 15 of the 17 alcohol-exposed subjects (1 missing from the D880 and T500 groups, due to technical error, Table 1). The average BAC for Group 5.25 g on PD 7 was 416.22 ± 9.93 mg/dl.

314

N.J. Murawski et al. / Behavioural Brain Research 236 (2013) 307–318

Fig. 4. Mean (±SEM) percentage of adaptive CR percentage (upper panel) and CR peak amplitude (lower panel) during long delay (left) and trace (right) eyeblink conditioning in undisturbed (UD), sham-intubated (SI), or alcohol-exposed (5.25 g) juvenile pups. Alcohol or sham intubations occurred on postnatal days (PD) 7–9 and 6 sessions of either long delay or trace eyeblink conditioning occurred on PD 26–28.

4.2.2. Performance measures – SR and UR amplitudes Mean (±SE) UMAs and SMAs can be found in Table 2. UMAs and SMAs did not differ across treatments for either long delay or trace conditioning tasks (Fs < 1.4), suggesting alcohol did not disrupt sensory, motivational, or motor functions necessary for conditioning performance.

4.2.3. Long delay conditioning Rats from Group 5.25 g showed significant long delay conditioning acquisition impairments compared to control groups (Fig. 4, left panels). Analysis of adaptive %CRs revealed a significant main effect of treatment [F(2,25) = 4.5, p < 0.03] without a sessions × treatment interaction (p > .4). Group 5.25 g also showed significantly reduced adaptive CMAs compared to control groups (ps < 0.02; Fig. 4 bottom left panel). Analysis of adaptive CMAs revealed a significant main effect of treatment [F(2,25) = 6.22, p < 0.01] and a significant sessions x treatment interaction [F(10,125) = 2.17, p < 0.03). Newman Keuls post hoc test revealed that Group 5.25 g showed significantly reduced adaptive CMAs during Sessions 5 and 6 compared to both Groups UD and SI (ps < 0.02).

4.2.4. Trace conditioning During trace eyeblink conditioning, rats from Group 5.25 g show significant reductions in adaptive %CRs compared to controls (ps < 0.01; Fig. 4 upper right panel). Analysis revealed a significant main effect of treatment [F(2,24) = 4.66, p < 0.01] without a sessions x treatment interaction (F < 0.6). Group 5.25 g also showed significantly lower adaptive CMAs compared to controls (Fig. 4 bottom right panel), with analysis revealing a significant main effect of treatment [F(2,24) = 5.32, p < 0.01] and a significant sessions × treatment interaction [F(10,120) = 2.88, p < 0.01]. Group 5.25 g showed significantly lower adaptive CMAs during Sessions 5 and 6 compared to both Groups UD and SI (ps < 0.05), which did not differ from one another (ps > 0.2). 4.3. Summary of experiment 2A In Experiment 2A, rats were exposed to a high binge dose of alcohol (5.25 g/kg/day) over PD 7–9, a period when Purkinje cells within the cerebellar cortex are more resilient to the neurotoxic effects of alcohol than at earlier points during the neonatal period and during which CA1 pyramidal cells remain vulnerable [14,15].

N.J. Murawski et al. / Behavioural Brain Research 236 (2013) 307–318

Contrary to our initial hypothesis, alcohol-exposed rats showed general deficits during eyeblink conditioning, with reductions in behavioral measures during both long delay and trace EBC relative to controls. 5. Experiment 2B: the effects of a moderately high binge dose of alcohol (4.66 g) over PD 7–9 on long delay and trace eyeblink conditioning Experiment 1 examined long delay and trace conditioning in juvenile exposed to both a high (5.25 g/kg/day; Exp. 1A) and moderate (4.66 g/kg/day; Exp. 1B) dose of alcohol over PD 4–9. In those experiments, eyeblink conditioning deficits in alcohol-exposed rats relative to controls were more robust at the higher alcohol dose. Here we further explore the effects of exposure to a moderate dose of alcohol (4.66 g/kg/day) limited to PD 7–9 on long delay and trace eyeblink conditioning. The primary question of interest was whether lower levels of alcohol exposure limited to PD 7–9 would differentially affect trace conditioning relative to long delay conditioning. 5.1. Materials and methods 5.1.1. Subjects The subjects were 100 Long Evans rats (50 males and 50 females) derived from 22 litters. 5.1.2. Design and procedures All alcohol dosing, surgery, design, experimental procedures and data analysis were identical to Experiment 2A, except that except a single binge alcohol dose of 4.66 g/kg/day (21.23% v/v) was administered. 5.2. Results Twenty-six animals were excluded from behavioral analyses (M = 12; F = 14). Nineteen animals showed excessively noisy EMG signals, including nine rats run in long delay (UD = 2; SI = 4; 4.66 g = 3) and 10 rats run in trace EBC (UD = 5; SI = 1; 4.66 g = 4). Three animals were dropped for low amplitude EMG recordings (URs ≤ 100 arbitrary units). An additional four animals were dropped due to experimenter error. The following analysis was performed on the remaining 74 rats, 38 of which were run on long delay (UD = 14, SI = 12, 4.66 g = 12) and 35 of which were run on trace EBC (UD = 9; SI = 12; 4.66 g = 14). 5.2.1. Body weights and BACs PD 7 and 9 body weights from four rats in Group UD were unavailable for analysis. Substantial weight gain was present in all groups between PD 7 and PD 9, although Group 4.66 g experienced moderate growth retardation during the intubation period (Table 1). These effects were supported by a significant main effect of days [F(1,63) = 645.88, p < 0.01] and a significant interaction of treatment x days [F(2,63) = 33.74, p < 0.01]. Newman-Keuls post hoc tests of these effects indicated no difference across treatment groups at PD 7 (ps > 0.5) with significant group differences at PD 9 – Groups UD and SI body weights did not differ from each other (p > .05), while Group 4.66 g body weights at PD 9 were only significantly lower than Group SI (p < 0.05). At PD 26, there were no treatment group differences in body weight (p > 0.72) although females weighed significantly less than males (p < 0.03; see Table 1). BACs were available for 14 of the 26 alcohol-exposed subjects. The average BAC for Group 4.66 g on PD 4 was 305.11 ± 22.61 mg/dl.

315

5.2.2. Performance measures – SR and UR amplitudes Mean (±SE) UR and startle amplitudes can be found in Table 2. Consistent with previous experiments, UR and SR amplitudes did not differ across treatment groups during long delay or trace conditioning (Fs < 2.9), again ruling out “performance effects” of alcohol. 5.2.3. Long delay conditioning Fig. 5 (upper left panel) shows adaptive %CRs as a function of treatment during long delay conditioning. There were no significant treatment or interaction effects evident on the adaptive %CR measure (Fs < 2.2). However, analysis of adaptive CMAs revealed a significant sessions × treatment interaction [F(10,175) = 3.58, p < 0.01), without a main effect of treatment (F < 2.3), that resulted from significantly lower CMAs in Group 4.66 g compared to controls during Session 6 (ps < 0.05; Fig. 4 lower left panel). 5.2.4. Trace conditioning Fig. 5 (upper right panel) shows adaptive %CRs as a function of treatment for trace conditioning. There was no significant effect of treatment or a sessions x treatment interaction (Fs < 1.1). Analysis of adaptive CMAs revealed a significant sessions × treatment interaction [F(10,160) = 2.68, p < 0.01], with Group SI showing elevated adaptive CMAs compared to both Groups UD and 4.66 g during Sessions 5 and 6 (ps < 0.05); Groups UD and 4.66 g did not differ across sessions (ps > 0.7; Fig. 5 lower right panel). 5.3. Summary of experiment 2B Experiment 2B examined long delay and trace eyeblink conditioning in juvenile rats exposed to a moderately high dose of alcohol (4.66 g/kg/day) over PD 7–9. Rats from Group 4.66 g showed some evidence of impairment during long delay eyeblink conditioning compared to controls in the adaptive CMA measure. However, during trace eyeblink conditioning, Group 4.66 g did not differ from Group UD on any measure. The significant increase in the adaptive CMA measure during trace conditioning in Group SI compared to Groups UD and 4.66 g is inconsistent with other experiments outlined above. 6. Meta-analysis of experiments 1A–2B To further examine the role of alcohol dose on eyeblink conditioning deficits we performed a “meta-analysis” that combined data from all four experiments, collapsing across the factors of task and window of exposure. ANOVA of adaptive CR percentage and amplitude revealed significant main effects of treatment. Mean (±SE) values for the UD, SI, 4.66 g, and 5.25 g groups, respectively were 71.23 (±1.39), 69.88 (±1.43), 59.76 (±2.33), 58.78 (±2.34), for adaptive %CR and 218.59 (±11.27), 208.24 (±10.48), 138.22 (±11.19), 121.90 (±8.47), for adaptive CMA. Post hoc tests revealed that both alcohol groups (5.25 g and 4.66 g) differed significantly from both control groups (UD and SI; ps < 0.01), without differing from each other (ps > 0.98). Controls also did not differ from one another (ps > 0.91). This analysis suggests that each alcohol dose produced a similar impairment in eyeblink conditioning. 7. Discussion The current findings reveal significant eyeblink conditioning impairments in juvenile rats exposed to alcohol over the neonatal period compared to controls. These findings are in general agreement with previous literature showing that juvenile and adult rats exposed to a high binge dose of alcohol (i.e., 5.25 g/kg/day) over the first postnatal week show deficits during short delay [13,44,47,48], long delay [48], and trace [31,49,50] eyeblink conditioning tasks.

316

N.J. Murawski et al. / Behavioural Brain Research 236 (2013) 307–318

Fig. 5. Mean (±SEM) adaptive CR percentage (upper panel) and CR peak amplitude (lower panel) during long delay (left) and trace (right) eyeblink conditioning in undisturbed (UD), sham-intubated (SI), or alcohol-exposed (4.66 g) juvenile pups. Alcohol or sham intubations occurred on postnatal days (PD) 7–9 and 6 sessions of either long delay or trace eyeblink conditioning occurred on PD 26–28.

We demonstrate that behavioral deficits occur when rats receive a high dose of alcohol over postnatal days (PD) 4–9, and extend the literature by (1) demonstrating these eyeblink conditioning deficits in rats receiving a single daily dose of alcohol, (2) examining a shorter window of alcohol exposure (PD 7–9), (3) examining long delay and trace conditioning with matched interstimulus intervals (ISIs), and (4) examining dose-response effects in these tasks. Previous studies in human subjects with FASD [3] and in juvenile rats exposed to 5.25 g/kg/day of alcohol over PD 4–9 [31] report comparable trace and delay eyeblink conditioning impairments. We also demonstrate significant trace and long delay eyeblink conditioning impairments in juvenile rats exposed to 5.25 g/kg/day of alcohol over PD 4–9 and PD 7–9 (Exp. 1A and 2A). Long delay conditioning was also impaired in rats exposed to the lower alcohol dose when administered over PD 4–9 (Exp. 1B) and PD 7–9 (Exp. 2B). However, trace conditioning impairments were only evident in rats exposed to the higher alcohol dose at both PD 4–9 and PD 7–9 windows (Exps. 1A and 2A), but not when exposed to the lower alcohol dose (4.66 g/kg/day), regardless of window of exposure (Exps. 1B and 2B). Taken together, these data suggest that neonatal alcohol exposure has a greater impact on long delay than trace eyeblink conditioning in juvenile rats, especially at lower alcohol doses.

Alcohol exposure during the neonatal period in the rat (PD 4–9) results in significant cell loss in a number of brain areas associated with the eyeblink conditioning circuit. These areas include components of the circuit shared by both delay and trace conditioning (e.g., the interpositus nucleus [44,51,52]; the pontine nucleus and the inferior olive [52]) and in areas differentially engaged by these tasks (i.e., cerebellar cortex [including both Purkinje and granule cells] [15,30,44,52,53]; the hippocampus [14,16,46]; and the prefrontal cortex [54]). Specific anatomical areas appear more sensitive to the teratogenic effects of alcohol over narrow temporal windows of exposure. For example, the greatest cell loss in the deep cerebellar nuclei (including the interpositus nucleus), Purkinje cells, granule cells, and the inferior olive occurs with alcohol exposure prior to PD 7 [15,30,51,52,55], with the pontine nucleus vulnerable over PD 7–9 [52]. Pyramidal cells in the hippocampus (particularly within CA1) remain vulnerable to alcohol over the entire neonatal period (PD 4–9 and PD 7–9) [14,16,46,56]. The frontal and other cortical regions show a somewhat later window of vulnerability compared to the hippocampus, which peaks around PD 7 and diminishes by PD 14 [14]. In the current study we manipulated the timing of alcohol exposure and hypothesized that distinct windows of exposure would result in specific eyeblink conditioning task deficits. Specifically,

N.J. Murawski et al. / Behavioural Brain Research 236 (2013) 307–318

we expected that rats exposed to alcohol over PD 4–9 would show deficits on both long delay and trace conditioning because the underlying shared neural circuit and the forebrain (hippocampus and mPFC) are targeted during this exposure window. Our findings support this hypothesis. Juvenile rats exposed to a high bingedose of alcohol (5.25 g/kg/day) over PD 4–9 showed comparable impairments on both long delay and trace eyeblink conditioning tasks. These results agree with the consistent finding that developmental alcohol exposure impairs eyeblink conditioning in both rats [13,17,31,36,44,47–49] and in human subjects with fetal alcohol spectrum disorders (FASD) [3,4], and extend these findings to include comparisons of delay and trace conditioning with matched interstimulus intervals (ISIs). By limiting our alcohol exposure window to PD 7–9, we attempted to minimize alcohol’s damaging effects on the cerebellum and brainstem, while still targeting forebrain structures such as the hippocampus and prefrontal cortex. Again, rats exposed to our highest dose of alcohol (5.25 g/kg/day) over PD 7–9 demonstrate conditioning impairments on both long delay and trace eyeblink tasks. To our knowledge, this is the first report of eyeblink conditioning impairments in juvenile rats exposed to alcohol over PD 7–9 [35,57]. Alcohol exposure limited to PD 7–9 produces deficits on other tasks impaired by hippocampal insult, including contextual fear conditioning and spatial water maze [14,33,58,59], although such exposure was insufficient to impair juvenile rats on a trace fear conditioning task [60]. Because PD 7–9 alcohol exposure did not differentially affect conditioning on long delay and trace conditioning, it is difficult with the current data set to draw firm conclusions about the underlying nature of these conditioning impairments. Although behavioral impairments are demonstrated in both long delay and trace conditioning, the affected systems underlying these behaviors may differ. While the cerebellum is most sensitive to alcohol prior to PD 7, in one study rats exposed to alcohol over PD 8–9 showed modest, but significant Purkinje cell reductions compared to control rats [30]. Our PD 7–9 dosing protocol may have resulted in cerebellar effects and thus may have contributed to impairments on the long delay conditioning task. In contrast, impairments in hippocampal function may account for some of the trace conditioning deficits that we observe [25,31]. It is also possible that impairments on both tasks result from a common source (e.g., alcohol’s effect on the pontine nucleus over the PD 7–9 window). Anatomical studies correlating cerebellar and hippocampal neuron number with behavioral performance on long delay and trace conditioning, respectively, in alcohol-exposed rats may shed light on this issue. In addition to the timing of exposure, another factor contributing to the teratogenicity of developmental alcohol exposure is the blood alcohol content (BAC) reached by a given dose [45]. BACs at or above ∼180 mg/dl are sufficient to cause significant Purkinje cell loss following PD 4–9 alcohol exposure [61], while BACs of ∼318 mg/dl are sufficient to cause significant CA1 pyramidal cell reductions within the hippocampus [45]. Impairments on standard delay eyeblink conditioning tasks require high BACs (∼380 mg/dl), although BACs as low as ∼160 mg/dl are sufficient to disrupt conditioning during particularly demanding multiple cue eyeblink conditioning tasks [17,36]. The two alcohol doses administered in our current study resulted in BACs of ∼400 mg/dl (the high dose) and ∼325 mg/dl (the low dose). BACs obtained through both of our chosen alcohol doses are sufficient to cause both Purkinje cell and CA1 pyramidal cell loss [45,61], and would likely disrupt behavior on both long delay and trace conditioning tasks when administered over the PD 4–9 period. However, we report impairments only during long delay conditioning at this dose. As previously mentioned, delay and trace conditioning with matched ISIs have a similar developmental trajectory [27,28], but differ in their sensitivity to hippocampal insult [25] and cholinergic manipulations

317

[62]. The lack of impairments during trace eyeblink conditioning in rats exposed to the lower alcohol dose may reflect a lack of alcohol targeting of the structures and/or systems mediating trace conditioning or perhaps compensatory actions by these or other structures that would affect trace, but not long delay, conditioning. The behavioral expression and underlying neural circuitry of eyeblink conditioning is conserved across a number of species allowing direct comparisons between rodent models of fetal alcohol spectrum disorders (FASD) and children with FASD [63]. Our findings of both trace and delay eyeblink conditioning deficits in juvenile rats exposed to a high dose of alcohol during the “brain growth spurt” are in agreement with recent findings in schoolaged children with known prenatal alcohol exposure [3]. Recent studies have utilized eyeblink conditioning to examine a variety of interventions and treatments in an effort to mitigate learning and memory deficits resulting from developmental alcohol exposure [31,50,64]. To further examine the contributions of cerebellar vs. hippocampal disturbances to performance on delay vs. trace eyeblink conditioning, respectively, future experiments utilizing anatomically targeted interventions will contribute to our understanding of the mechanisms behind the learning and memory deficits observed in FASD. Acknowledgements The authors would like to thank Felipe L. Schiffino and Henry S. Lange for technical support. This work was supported by the University of Delaware and by NIH grant RO1 AA014288-01 to MES. References [1] Spadoni AD, Bazinet AD, Fryer SL, Tapert SF, Mattson SN, Riley EP. BOLD response during spatial working memory in youth with heavy prenatal alcohol exposure. Alcoholism, Clinical and Experimental Research 2009;33:2067–76. [2] Kodituwakku PW. Neurocognitive profile in children with fetal alcohol spectrum disorders. Developmental Disabilities Research Reviews 2009;15:218–24. [3] Jacobson SW, Stanton ME, Dodge NC, Pienaar M, Fuller DS, Molteno CD, et al. Impaired delay and trace eyeblink conditioning in school-age children with fetal alcohol syndrome. Alcoholism, Clinical and Experimental Research 2011;35:250–64. [4] Jacobson SW, Stanton ME, Molteno CD, Burden MJ, Fuller DS, Hoyme HE, et al. Impaired eyeblink conditioning in children with fetal alcohol syndrome. Alcoholism, Clinical and Experimental Research 2008;32:365–72. [5] Uecker A, Nadel L. Spatial locations gone awry: object and spatial memory deficits in children with fetal alcohol syndrome. Neuropsychologia 1996;34:209–23. [6] Hamilton DA, Kodituwakku P, Sutherland RJ, Savage DD. Children with fetal alcohol syndrome are impaired at place learning but not cued-navigation in a virtual Morris water task. Behavioural Brain Research 2003;143:85–94. [7] Willoughby KA, Sheard ED, Nash K, Rovet J. Effects of prenatal alcohol exposure on hippocampal volume, verbal learning, and verbal and spatial recall in late childhood. Journal of International Neuropsychological Society 2008;14:1022–33. [8] Mattson SN, Riley EP. A review of the neurobehavioral deficits in children with fetal alcohol syndrome or prenatal exposure to alcohol. Alcoholism, Clinical and Experimental Research 1998;22:279–94. [9] Norman AL, Crocker N, Mattson SN, Riley EP. Neuroimaging and fetal alcohol spectrum disorders. Developmental Disabilities Research Reviews 2009;15:209–17. [10] Spadoni AD, McGee CL, Fryer SL, Riley EP. Neuroimaging and fetal alcohol spectrum disorders. Neuroscience and Biobehavioral Reviews 2007;31:239–45. [11] Riikonen R, Salonen I, Partanen K, Verho S. Brain perfusion SPECT and MRI in foetal alcohol syndrome. Developmental Medicine and Child Neurology 1999;41:652–9. [12] Coles CD, Goldstein FC, Lynch ME, Chen X, Kable JA, Johnson KC, et al. Memory and brain volume in adults prenatally exposed to alcohol. Brain and Cognition 2011;75:67–77. [13] Stanton ME, Goodlett CR. Neonatal ethanol exposure impairs eyeblink conditioning in weanling rats. Alcoholism, Clinical and Experimental Research 1998;22:270–5. [14] Marino MD, Aksenov MY, Kelly SJ. Vitamin E protects against alcohol-induced cell loss and oxidative stress in the neonatal rat hippocampus. International Journal of Developmental Neuroscience 2004;22:363–77.

318

N.J. Murawski et al. / Behavioural Brain Research 236 (2013) 307–318

[15] Hamre KM, West JR. The effects of the timing of ethanol exposure during the brain growth spurt on the number of cerebellar Purkinje and granule cell nuclear profiles. Alcoholism, Clinical and Experimental Research 1993;17:610–22. [16] Tran TD, Kelly SJ. Critical periods for ethanol-induced cell loss in the hippocampal formation. Neurotoxicology and Teratology 2003;25:519–28. [17] Brown KL, Calizo LH, Stanton ME. Dose-dependent deficits in dual interstimulus interval classical eyeblink conditioning tasks following neonatal binge alcohol exposure in rats. Alcoholism, Clinical and Experimental Research 2008;32:277–93. [18] Gil-Mohapel J, Boehme F, Kainer L, Christie BR. Hippocampal cell loss and neurogenesis after fetal alcohol exposure: insights from different rodent models. Brain Research Reviews 2010;64:283–303. [19] Thompson RF. In search of memory traces. Annual Review of Psychology 2005;56:1–23. [20] Thompson RF, Steinmetz JE. The role of the cerebellum in classical conditioning of discrete behavioral responses. Neuroscience 2009;162:732–55. [21] Gerwig M, Kolb FP, Timmann D. The involvement of the human cerebellum in eyeblink conditioning. Cerebellum 2007;6:38–57. [22] Skelton RW. Bilateral cerebellar lesions disrupt conditioned eyelid responses in unrestrained rats. Behavioral Neuroscience 1988;102:586–90. [23] Woodruff-Pak DS, Steinmetz JE. Eyeblink classical conditioning. Applications in Humans, vol. 1. Boston: Kluwer Academic Publishers; 2000. [24] Woodruff-Pak DS, Steinmetz JE. Classical eyeblink conditioning. Animal Models, vol. 2. Boston: Kluwer Academic Publishers; 2000. [25] Ivkovich D, Stanton ME. Effects of early hippocampal lesions on trace, delay, and long-delay eyeblink conditioning in developing rats. Neurobiology of Learning and Memory 2001;76:426–46. [26] Clark RE, Squire LR. Classical conditioning and brain systems: the role of awareness. Science 1998;280:77–81. [27] Barnet RC, Hunt PS. Trace and long-delay fear conditioning in the developing rat. Learning and Behavior 2005;33:437–43. [28] Ivkovich D, Paczkowski CM, Stanton ME. Ontogeny of delay versus trace eyeblink conditioning in the rat. Developmental Psychobiology 2000;36:148–60. [29] Dobbing J, Sands J. Comparative aspects of the brain growth spurt. Early Human Development 1979;3:79–83. [30] Thomas JD, Goodlett CR, West JR. Alcohol-induced Purkinje cell loss depends on developmental timing of alcohol exposure and correlates with motor performance. Brain Research Developmental Brain Research 1998;105: 159–66. [31] Thomas JD, Tran TD. Choline supplementation mitigates trace, but not delay, eyeblink conditioning deficits in rats exposed to alcohol during development. Hippocampus 2012;22(3):619–30. [32] Murawski NJ, Stanton ME. Variants of contextual fear conditioning are differentially impaired in the juvenile rat by binge ethanol exposure on postnatal days 4–9. Behavioural Brain Research 2010;212:133–42. [33] Murawski NJ, Stanton ME. Effects of dose and period of neonatal alcohol exposure on the context preexposure facilitation effect. Alcoholism, Clinical and Experimental Research 2011;35(6):1160–70. [34] Kelly SJ, Lawrence CR. Intragastric intubation of alcohol during the perinatal period. Methods in Molecular Biology 2008;447:101–10. [35] Brown KL, Calizo LH, Goodlett CR, Stanton ME. Neonatal alcohol exposure impairs acquisition of eyeblink conditioned responses during discrimination learning and reversal in weanling rats. Developmental Psychobiology 2007;49:243–57. [36] Brown KL, Burman MA, Duong HB, Stanton ME. Neonatal binge alcohol exposure produces dose dependent deficits in interstimulus interval discrimination eyeblink conditioning in juvenile rats. Brain Research 2009;1248: 162–75. [37] Stanton ME, Freeman Jr JH. Eyeblink conditioning in the infant rat: an animal model of learning in developmental neurotoxicology. Environmental Health Perspectives 1994;102(Suppl. 2):131–9. [38] Hodes GE, Shors TJ. Distinctive stress effects on learning during puberty. Hormones and Behavior 2005;48:163–71. [39] Stanton ME, Tran DT, Robinette BL, Goodlett CR. Neonatal Alcohol Exposure Impairs Eyeblink Conditioning in Weanling Rats: Dose-Response Effects. New Orleans, LA: Society for Neuroscience; 2000. [40] Plakke B, Freeman JH, Poremba A. Metabolic mapping of the rat cerebellum during delay and trace eyeblink conditioning. Neurobiology of Learning and Memory 2007;88:11–8. [41] Brown KL, Agelan A, Woodruff-Pak DS. Unimpaired trace classical eyeblink conditioning in Purkinje cell degeneration (pcd) mutant mice. Neurobiology of Learning and Memory 2010;93:303–11.

[42] Plakke B, Freeman JH, Poremba A. Metabolic mapping of rat forebrain and midbrain during delay and trace eyeblink conditioning. Neurobiology of Learning and Memory 2009;92:335–44. [43] Woodruff-Pak DS, Disterhoft JF. Where is the trace in trace conditioning? Trends in Neurosciences 2008;31:105–12. [44] Tran TD, Jackson HD, Horn KH, Goodlett CR. Vitamin E does not protect against neonatal ethanol-induced cerebellar damage or deficits in eyeblink classical conditioning in rats. Alcoholism, Clinical and Experimental Research 2005;29:117–29. [45] Bonthius DJ, West JR. Permanent neuronal deficits in rats exposed to alcohol during the brain growth spurt. Teratology 1991;44:147–63. [46] Livy DJ, Miller EK, Maier SE, West JR. Fetal alcohol exposure and temporal vulnerability: effects of binge-like alcohol exposure on the developing rat hippocampus. Neurotoxicology and Teratology 2003;25:447–58. [47] Green JT, Johnson TB, Goodlett CR, Steinmetz JE. Eyeblink classical conditioning and interpositus nucleus activity are disrupted in adult rats exposed to ethanol as neonates. Learning and Memory 2002;9:304–20. [48] Tran TD, Stanton ME, Goodlett CR. Binge-like ethanol exposure during the early postnatal period impairs eyeblink conditioning at short and long CS-US intervals in rats. Developmental Psychobiology 2007;49:589–605. [49] Tran DT, Goodlett CR. Dose-related effects of binge-like ethanol exposure during the early neonatal period on trace eyeblink conditioning in adult rats. Society for Neuroscience. San Diego: Neuroscience Meeting Planner (online); 2004. p. Program No. 326.12. [50] Schiffino FL, Jablonski SA, Hamilton GF, St. Cyr SA, Finamore JM, Greenough WT, et al. Voluntary exercise followed by environmental complexity reverses deficits in trace eyeblink conditioning and adult hippocampal neurogenesis in a rat model of fetal alcohol spectrum disorder. Baltimore, MD: Pavlovian Society [abstract]. [51] Young BW, Sengelaub DR, Steinmetz JE. MK-801 administration during neonatal ethanol withdrawal attenuates interpositus cell loss and juvenile eyeblink conditioning deficits. Alcohol 2010;44:359–69. [52] Dikranian K, Qin YQ, Labruyere J, Nemmers B, Olney JW. Ethanol-induced neuroapoptosis in the developing rodent cerebellum and related brain stem structures. Brain Research Developmental Brain Research 2005;155:1–13. [53] Lee Y, Rowe J, Eskue K, West JR, Maier SE. Alcohol exposure on postnatal day 5 induces Purkinje cell loss and evidence of Purkinje cell degradation in lobule I of rat cerebellum. Alcohol 2008;42:295–302. [54] Whitcher LT, Klintsova AY. Postnatal binge-like alcohol exposure reduces spine density without affecting dendritic morphology in rat mPFC. Synapse 2008;62:566–73. [55] Green JT, Arenos JD, Dillon CJ. The effects of moderate neonatal ethanol exposure on eyeblink conditioning and deep cerebellar nuclei neuron numbers in the rat. Alcohol 2006;39:135–50. [56] Ikonomidou C, Bittigau P, Ishimaru MJ, Wozniak DF, Koch C, Genz K, et al. Ethanol-induced apoptotic neurodegeneration and fetal alcohol syndrome. Science 2000;287:1056–60. [57] Johnson TB, Tran DT, Goodlett CR. Eyeblink classical conditioning deficits in adult rats induced by 3-day neonatal binge alcohole exposure. Alcoholism, Clinical and Experimental Research 2003;26 (Supplement). [58] Goodlett CR, Johnson TB. Neonatal binge ethanol exposure using intubation: timing and dose effects on place learning. Neurotoxicology and Teratology 1997;19:435–46. [59] Johnson TB, Goodlett CR. Selective and enduring deficits in spatial learning after limited neonatal binge alcohol exposure in male rats. Alcoholism, Clinical and Experimental Research 2002;26:83–93. [60] Hunt PS, Jacobson SE, Torok EJ. Deficits in trace fear conditioning in a rat model of fetal alcohol exposure: dose-response and timing effects. Alcohol 2009;43:465–74. [61] Green JT. The effects of ethanol on the developing cerebellum and eyeblink classical conditioning. Cerebellum 2004;3:178–87. [62] Hunt PS, Barnet RC, Shea ME, Baker EM. Cholinergic modulation of trace conditioning trained in serial compound: a developmental analysis. Neurobiology of Learning and Memory 2006;86:311–21. [63] Goodlett CR, Stanton ME, Steinmetz JE. Alcohol-induced damage to the developing bra functional approaches using classical eyeblink conditioning. In: Woodruff-Pak DS, Steinmetz JE, editors. Eyeblink Classical Conditioning. Animal Models, vol. 2. Boston/Dordrecht/London: Kluwer Academic Publishers; 2000. p. 134–53. [64] Wagner JL, Garzon D, McKay J, Klintsova AY, Greenough WT, Goodlett CR. Complex motor learning ameliorates deficits in eyeblink classical conditioning in a rat model of binge alcohol exposure during the third trimester equivalent. Developmental Psychobiology 2008;50:748.