Auditory Deficits in a Mouse Model of First-Trimester Prenatal Alcohol Exposure

Prenatal alcohol exposure (PAE) may lead to several problems including fetal alcohol syndrome (FAS), partial FAS, alcohol-related neurodevelopmental disorders, and alcohol-related birth defects, collectively called fetal alcohol spectrum disorders (FASDs) [1]. A magnetic resonance imaging study of ∼1,000 children with FASD identified PAE as the dominant risk factor, among 14 other prenatal and postnatal risks, for interpreting variations in regional brain size and brain function [2]. Thus, it is important to understand how PAE affects brain development.

A strong presentation of auditory deficits has been established in PAE individuals with characterized symptoms of FAS, the most severe form of FASD. FAS is characterized with facial anomalies, growth delays, and neurologic problems that span from hyperactivity, attention deficits, impulsiveness, anxiety, social difficulties, and language delay to intellectual disability. Audiometry tests revealed a high prevalence (30–70%) of conductive and sensorineural hearing loss in FAS children and adolescents [3, 4]. Additionally, children with FAS show impaired central auditory function, including abnormal speech perception in noise, a binaural processing deficit, and reduced incidence of the expected “right ear advantage” on dichotic speech perception tests [5, 6]. A study of infants whose mothers had evidence of alcohol abuse during pregnancy reported abnormal auditory brainstem responses (ABRs) [7], an objective test to evaluate the integrity of the auditory system from the cochlea to the brainstem. Moreover, cortical event-related potentials (ERPs) in children (aged 11–15 years) with FAS or partial FAS demonstrated abnormal cortical activity when performing an auditory task [8].

Auditory dysfunction was also detected in a general population with PAE, regardless of whether the individuals were diagnosed with FAS or not. Cortical ERPs in school-age children (4–18 years) with FASD displayed prolonged peak latencies and/or reduced peak amplitudes [9, 10]. A group of prenatal alcohol-exposed adolescents (aged 13–14 years) who did not have phenotypic changes observed in FAS showed altered cortical ERP components and a higher incidence of hearing loss and abnormal auditory processing [11, 12]. Delayed auditory responses (prolonged peak latencies) were also detected by magnetoencephalography in young FASD children (aged 3–6 years) with normal hearing sensitivity [13]. Additionally, adults with FASD performed worse with auditory stimuli than with visual cues on an attention task [5, 14]. These studies support the effect of PAE on auditory function at ages when the auditory system is developed. A recent study of ∼1,000 newborn and 1-month-old infants, however, reported no effect of PAE on ABR measures [15]. This highlights the complexity of PAE’s effects on auditory development and how these effects depend on the age and pattern of alcohol exposure. To untangle this complexity, it is essential to characterize the developmental trajectory of auditory function and its relationship with other comorbid symptoms following PAE during different developmental periods and at varying degrees. This approach is crucial for establishing a mechanistic understanding of FASD neuropathology, particularly given the increasing appreciation of auditory dysfunction in affecting metalinguistic skills and overall academic performance [16].

Several animal models of PAE provide evidence of peripheral and central auditory abnormalities. For example, PAE induced excessive cell death in the mouse embryonic otic vesicle [17]. Prenatal alcohol-exposed rats showed a high incidence of sensorineural hearing loss, hair cell lesions and stereocilia malformation, and elevated ABR thresholds and prolonged latencies in post-weanlings and adults [18‒21]. At the cellular level, auditory brainstem neurons responsible for binaural processing had shorter dendrites and abnormal evoked neuronal activity in avian embryos [7] and reduced neuronal number in primates [22]. These studies established the relevance of animal models for studying PAE-associated auditory deficits.

This study aimed to determine changes in auditory function in mice at both developing and mature ages following a PAE paradigm during the first half of embryonic development from gestational day 1 (G1) to G10, mimicking alcohol consumption during the first trimester of pregnancy in humans. We used ABRs as a functional test of auditory system development at the age of hearing onset (postnatal day 14 or P14 in mice) and in young adulthood (2 months). Additionally, we examined the effect of PAE on the auditory gating function using acoustic startle responses (ASRs) and prepulse inhibition (PPI) as behavioral readouts. ASR is an unconditional reflex in response to an intense and sudden sound stimulus. Previous studies showed that alcohol exposure throughout gestation resulted in altered acoustic startle and PPI in adult rats and monkeys as well as human fetuses [23‒25]. Besides FASD, alterations in ABR and PPI measures have been linked to various neurodevelopmental disorders that exhibit auditory problems [26‒29], highlighting the importance of studying these two tests to understand FASD. Furthermore, both assays are widely used in animal models and human individuals and therefore have the potential for direct clinical implications. Finally, locomotor activity and anxiety-like behavior were assessed using the open field test (OFT) and elevated zero maze (EZM) tests to compare PAE and control offspring. We chose these assays because previous studies reported anxiety-related behavioral changes and/or hyperactivity in rodents following long-term PAE [30, 31]. The goal of this study was to identify the specific deficits related to PAE that are associated with alcohol consumption during the early stages of gestation.

Mouse breeders on the FVB background (inbred laboratory mouse strain named after its susceptibility to Friend leukemia virus B) were obtained from Jackson Laboratories and used to establish colonies at Florida State University. Mice were housed under an inverted light cycle (12-h light/dark cycle). All animal procedures were approved by the Florida State University Institutional Animal Care and Use Committee and performed in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals.

Breeder mice were group-housed separated by sex, fed on standard rodent chow (LabDiet 5015), and had water ad libitum before reaching 7 weeks of age. At this point, female mice were individually housed and fed a Lieber-DeCarli liquid diet (Dyets D710260) containing maltose dextrin for 1 week to acclimate. When the female mice reached 8 weeks of age, they were paired with males. Females were separated from males the following day. This day was designated as the first day of gestation (G1). Next, the female mice were randomly assigned to either the control group or the alcohol-exposed group (Fig. 1a). In the control group, females continued consuming the liquid diet (alcohol-free) for 10 days, from G1 to G10. For the alcohol-exposed group, ethanol was gradually added to the diet, reaching a maximum concentration of 32% kcal. Specifically, on day 1 (G1), mice received a diet with 8% kcal from ethanol, which increased over subsequent days to 16%, 24%, 28%, and finally 32% kcal. The 32% kcal diet was continued for 3 days, from G5–G7. Afterward, the mice were gradually weaned off alcohol by consuming diets with respective ethanol concentrations of 24%, 16%, and 8% kcal in the following days. On G11, all female mice (both control and alcohol-exposed) were returned to standard chow (LabDiet 5015) and continuously monitored until pup birth. Blood alcohol concentration (BAC) was measured in alcohol-exposed dams on G7 (the third day at peak ethanol concentration) and averaged 119.75 mg/dL in alcohol-exposed females across three cohorts.

The day of pup birth was defined as postnatal day 0 (P0). Both the control and alcohol-exposed groups had large litter sizes (ranging from 7 to 12 pups). Each litter was culled to 6–8 pups, with consideration given to ensuring an equal number of males and females, if possible. Offspring were tested for ABRs and ASRs at 2 weeks (P14) and 2 months (P60–P70) of age. At a later age, the OFT and EZM were also performed to evaluate anxiety-like behaviors. Dams were only used one time for breeding and were excluded after one exposure to liquid diet.

ABRs were recorded at P14 and 2 months of age as previously described [32]. Briefly, animals were anesthetized with an intraperitoneal injection of 105 mg/kg ketamine and 9 mg/kg xylazine. Mice were placed on a heating pad to maintain their body temperature during ABR recordings. Recording was conducted in a sound-attenuating chamber using an ABR acquisition system (Tucker Davis Technologies; Alachua, FL, USA). Subdermal needle electrodes (Rochester Electro-Medical; Lutz, FL, USA) were used for recording, with the positive electrode positioned at the vertex of the skull, the reference electrode below the pinna of the right ear, and the ground electrode in the right thigh. A multi-field speaker (MF1; Tucker Davis Technologies) was placed 10 cm from the right ear for open-field recording. Click (0.1 ms in duration) and tone pip bursts of 4, 8, 16, 24, and 32 kHz (5 ms in duration; gate time: 2 ms to reach within 3 dB of the peak stimuli; gate type: cosine) were presented at a rate of 21/s. Sound level calibration for clicks was performed using a 1/4″ free-field measure calibration mic kit (PCB-378C01; PCB Piezotronics; Depew, NY, USA) using the RZ6_Click_Calibration.rcx circuit (Tucker Davis Technologies). Tone calibration was performed using BioSigRZ software (Tucker Davis Technologies) by conducting a sweep over the frequency range of interest and generating a speaker response curve along with a correction curve that would be required to flatten the response. A finite impulse response filter was calculated from the correction curve and used to generate a flat response across all frequencies. The finite impulse response filter was validated monthly during the duration of the project for consistency. Each type of stimuli (click and each frequency of tone) was presented at levels from 90-dB sound pressure level (SPL) to 20 dB SPL first in 10-dB decrements and then repeated in 5-dB decrements at levels near the threshold. Biological responses were pre-amplified (100×; Mdeusa4Z Pre-amp/Digitizer and Headstage; sampling rate of 12 kHz; Tucker Davis Technologies), bandpass filtered (300–3,000 Hz with 60 Hz notch filter), digitized (RZ6-A-P1 bioacoustic system; Tucker Davis Technologies), and averaged from a total of 256 trials over a 20-ms window using BioSigRZ software (Tucker Davis Technologies).

Both click stimulus and tone bursts were used to identify changes in ABR thresholds and peak latencies. ABR threshold was defined as the lowest SPL that evokes a detectable and repeatable response. If no ABR wave was detected at 90 dB for a specific stimulus, a nominal threshold of 95 dB was assigned. No ABRs at 90 dB were only found at the age of P14 for tone stimuli. This includes 3 control and 3 PAE cases in response to 4 kHz tones and 2 PAE cases in response to 32 kHz tones. In this study, the detected ABR thresholds in response to click and most sensitive frequencies (16–24 kHz) range from 15 to 40 dB SPL at the age of 2 months, consistent with previous reports [33, 34] and validating the sensitivity of our ABR setup.

Peak latencies were measured for the first four peaks to click stimulus and tone bursts at 90 dB SPL, as described previously [32]. At P14, waves I and II were readily identifiable, often followed by a broad wave at the approximate location of adult waves III and IV. ABR size was evaluated with several measures made from ABR waveforms to click stimulus at 90 dB SPL. The peak amplitude of ABR wave I was taken from the peak of wave I to the trough that immediately followed it. The ABR integral was determined by summing the area under the baseline during a period of 0.82–7.39 ms from click onset, which contains waves I–V at P14 and waves I–IV at 2 months of age. ABR integral is the sum of the magnitudes of the waveform (not just the peak magnitudes), regardless of direction, which provides a measure of the overall signal strength. These measures were statistically compared between control and PAE offspring. Additionally, the ABR waveforms were averaged within each group and superimposed to illustrate the differences between groups. This comparison was performed using either the original waveforms or after being temporally adjusted to the peak of wave II to better reflect changes in the waves generated in the brain.

ASRs were measured at two ages, P14 and 2 months, using an SR-LAB Startle Response System (San Diego Instruments, San Diego, CA, USA). A single mouse was placed in the Plexiglas enclosure and acclimated for 5 min. After acclimation, the mouse was exposed to two consecutive test procedures. In each procedure, the stimulation consisted of broadband startle pulses (115 dB; 40 ms duration) either alone (n = 24 trials) or with a prepulse (20 ms duration) at 68 dB, 75 dB, or 80 dB SPL (n = 8 trials per prepulse level). The gap between the prepulse and startle pulse was 80 ms. All trials (total 48) were arranged in a pseudorandom order with randomly assigned intertrial intervals between 5 and 25 s. The two procedures contained the same trials, each with a different pseudorandom order and randomly assigned intertrial intervals. Together, the dataset collected from each mouse contained ASRs at 115 dB (48 trials), 115 dB with 68-dB prepulse (16 trials), 115 dB with 75-dB prepulse (16 trials), 115 dB with 80-dB prepulse (16 trials). Throughout the period of acclimation and recording, a continuous background noise at 65 dB SPL was presented.

For each trial, ASR was collected starting 40 ms before the onset of the startle pulse (115 dB) for a total of 200 ms. ASR parameters including the latency and peak amplitude of the startle responses were measured and averaged across the same type of trials. The mean activity during the 40 ms prior to the presence of the startle pulse was used as the baseline floor. Startle responses were measured as the highest activity amplitude (mV Max) within the recording time window of 40–90 ms. Startle latency was measured as the time when the peak amplitude was present from the onset of the startle pulse. PPI with 68-dB prepulse was calculated following the formula: PPI = 1−(mean mV Max_{68 dB prepulse}/mean mV Max_{no prepulse}). The same formula applied to the calculation of PPIs with 75-dB and 80-dB prepulses.

Locomotor activity and anxiety-like behavior were assessed using the OFT in 2-month-old mice as previously described [35]. Mice were acclimated to the testing room for at least 30 min before the test began. An individual mouse was placed into a square 18-inch diameter Plexiglas box located inside of a sound-attenuating room. The room was lit with a red light, and white noise (approximately 65 dB) served as the background sound. The activity of the mouse was tracked for 10 min using AnyMaze software (Wood Dale, IL, USA). For data analysis, the arena was divided into nine sections including one large center section, four corner sections, and four surrounding sections. The total distance traveled as well as the distance traveled and time spent in each section were measured for each mouse. Data were also collected and analyzed for the first and last 5 min, in addition to the total 10-min period, to identify the potential influence of a novice environment.

EZM was used as an additional test to assess anxiety-like behavior in 2-month-old mice as previously described [35]. Mice were acclimated to the testing room for at least 30 min before the test began. An individual mouse was placed in an open arm of the maze located inside of a sound-attenuating room. The room was lit with a red light, and white noise (approximately 65 dB) served as the background sound. The maze was continuously videotaped with a camera for a total of 5 min. Mice that fell off the maze before the end of the 5 min were allowed to rest for at least 30 min before retesting. Mice that failed two consecutive tests were removed from the study. Of the 35 mice tested, 3 mice failed the EZM testing and were removed from the study. Video data were analyzed by two blinded investigators. First, total arm entry was measured as the number of times that the mouse alternated between the open and closed arms of the maze. An inclusion criterion of 10 total arms was used. Next, latency to closed arm was measured as the time that a mouse took to enter a closed arm of the maze upon its initial placement on the apparatus. Finally, total time in open arms was calculated as the percentage of time that a mouse spent in the open arms of the maze per the total test time.

To control for environmental factors that may change over time, three out of six completed breeding cycles that generated offspring from both control and alcohol-exposed dams were used in this study. From the three cohorts, 6 control dams and 7 alcohol-exposed dams were included. Each of the six analyses (ABR and ASR at P14; ABR, ASR, EZM, and OFT at 2 months of age) was performed on offspring derived from at least 3 control and 3 alcohol-exposed dams. The sample sizes tested for each analysis are provided in Table 1. No more than 4 animals per litter were used for each analysis. For P14 tests, two separate groups of mice underwent either ABR or ASR recording, but not both, to avoid confounding influence between the two auditory-associated tests. At 2 months, all animals underwent tests between P60 and P70, with a 2-day interval between tests, in the same order of OFT, EZM, ABR, and ASR. The sample sizes varied slightly across the four tests at this age because we performed behavioral tests within a fixed time window on the test day to minimize potential influence of circadian rhythm. The investigator was blinded to the experimental group identity during the tests and measures. Each individual animal served as a data point for statistical analyses, with samples grouped by age and separated by sex.

A series of mixed model ANOVAs followed by post hoc contrasts, where appropriate, were used to analyze body weight, auditory functional measures (ABR threshold, size, and latency), acoustic-driven behavior (ASR amplitude and PPI), OFT, and EZM. Statistical analyses were performed using SAS/STAT 9.4 (SAS, Carey, NC, USA). Litter was included as a random factor. Treatment and sex were used as between-subjects factors. Additionally, ABR peak latency was analyzed with wave as a within-subject factor. PPI was analyzed separately for each prepulse level. For illustration in the figures, data from male and female mice for each group were separated when a significant effect of sex was observed; otherwise, combined. Post hoc contrast analyses were performed only when a statistically significant interaction between treatment and sex was found. Significance was determined by p < 0.05. Graphs were prepared using Prism (version 10.1.2; GraphPad Prism, San Diego, CA, USA).

Offspring born from the control and alcohol-exposed dams are named control and PAE mice, respectively, in the subsequent description, tables, and figures. The PAE exposure paradigm did not affect the overall growth of the offspring (Fig. 1b). At P14, the data did not reveal significant effects of treatment (F_(1,36) = 0.08; p = 0.7765), sex (F_(1,36) = 0.25; p = 0.6194) or treatment × sex interaction (F_(1,36) = 1.45; p = 0.2367) on bodyweight. Similarly, at 2 months, there were no significant effects of treatment (F_(1,25) = 1.24; p = 0.2755) or treatment × sex interaction (F_(1,25) = 0.00; p = 0.9819) on body weight. As expected, there was an effect of sex (F_(1,25) = 42.52; p < 0.001), reflecting higher bodyweight in males than females.

Both control and PAE offspring displayed a developmental decline of ABR thresholds in response to clicks and tones, reflecting auditory maturation during development (Fig. 2, 3). There was no significant effect of treatment, sex, or treatment × sex interaction at P14 or 2 months, except for an interaction between treatment and sex at 2 months in response to 32 kHz tone (Table 2). Similarly, both control and PAE offspring displayed a developmental decline of ABR peak latencies, which were not affected by treatment, sex, or treatment × sex interaction at either age (Fig. 4a, b; Table 2). PAE had no effect on interpeak intervals, measured from peak I–II and II–V at P14 and peak I–II and II–III at 2-month-old. Interpeak intervals involving waves III–V at 2-month-old were not consistently identified across cases (see below) and thus not analyzed further.

There was no effect of treatment or treatment × sex interaction on the wave I peak amplitude at either age (Table 2; Fig. 5a–c). As expected [36‒41], there was an effect of sex on wave I amplitude at 2 months, reflecting a larger amplitude in females than in males. The ABR integral at 2 months was significantly affected by treatment and sex, demonstrating larger ABR sizes in females and reduced ABR sizes following PAE treatment (Fig. 5d–f; Table 2). The treatment × sex interaction was not significant, indicating that PAE affects ABR size similarly in males and females.

To further identify the potential site of PAE-induced ABR size reduction at the age of 2 months, we compared the average ABR waveforms across all individuals of the same group between control and PAE offspring. In the control group, waves I, II, and V were distinct in the average ABRs, while waves III and IV were merged as one wave, probably due to varying peak latencies at these sites across individuals (Fig. 6a). A reduction in the average ABR voltages was observed across the course of the ABR waves in the PAE group, with the most notable reductions observed at later waves III–IV. Although reduced in size, the ABR shape at waves I, II, and V were similar between control and PAE groups. The shape appeared different between groups during the time window between peaks II and V. To confirm whether this difference was generated artificially due to individual variations in peak latencies, we temporally realigned ABR waveforms to the onset of the wave II peak and identified a smaller wave III–IV in the averaged ABR waveform in the PAE group as compared to the control (Fig. 6b). Further examination of the ABR shapes of individual animals revealed that 86% (12 out of 14) of control offspring displayed distinct peaks III and IV (Fig. 6c, left column). During the same time window, only 71% (15 out of 21) of PAE offspring had two differentiable peaks, which were often less distinct from each other and had irregular intervals (Fig. 6c, right column). The averaged ABR waveform at P14, after realignment to the onset of the wave II peak, also displayed a smaller wave III–IV (Fig. 6d), raising the possibility that this deficit occurred early during development. Together, these analyses identified the brainstem areas responsible for generating ABR waves III–IV as a potential site of PAE impact.

There was no significant effect of either treatment, sex, or treatment × sex interaction on the peak ASR amplitude (VMax) at either P14 or 2 months of age (Table 3; Fig. 7a, b). Our analysis did reveal a significant effect of treatment (PAE) on 68-dB prepulse induced PPI at P14 (Table 3; Fig. 7c), reflecting reduced PPI by PAE at this age. Additionally, a higher percentage of mice in the PAE group failed to display an effective PPI (5% cutoff) to a 68-dB prepulse (37% in control vs. 67% in PAE) and a 75-dB prepulse (12% in control vs. 33% in PAE) at P14. Similarly, at age 2 months, a higher percentage of mice in the PAE group failed to display an effective PPI to a 68-dB prepulse (17% in control vs. 40% in PAE). Together, the ASR and PPI analyses demonstrate that PAE leads to a lower possibility of triggering effective PPI by low levels of prepulses at both developing and adult ages.

In the OFT, there was no significant effect of treatment, sex, or treatment × sex interaction on the total distance traveled (a measure of activity) or the time spent in the center of the arena (a measure of anxiety; Table 4; Fig. 8a, b). For the EZM, the total time in the open arms (a measure of anxiety) was not significantly affected by treatment, sex, or treatment × sex interaction (Table 4; Fig. 8c). PAE resulted in a reduced number of total arm entries (a measure of exploration; Fig. 8d).

In this study, we developed a mouse model of PAE during the first half of embryonic development, which is comparable to the first trimester of human fetal development. We did not identify a significant effect of first-trimester PAE on either ABR thresholds or ABR peak latencies in mice. This observation indicates that first-trimester PAE does not cause substantial damage to the overall hearing sensitivity in the inner ear or the signal transduction from the ear to the auditory brainstem and within the brainstem. Hearing preservation, despite early alcohol exposure, may be linked to the developmental trajectory of the auditory system. In mice, both hair cells and ganglion neurons arise from the otic placode. It is not until G9 that neuroblasts begin to delaminate from the otic vesicle to form the cochlea-vestibular ganglion (CVG) between G10 and 12 [42]. As a result, many key developmental events such as CVG separation into the spiral and vestibular portions, spiral ganglion neuron diversification into functional subtypes, and the establishment of the periphery and central projections occur after the cessation of alcohol intake in the current PAE model. Similarly, hair cell development in the cochlea begins around G14 in mice [43‒45]. Even though auditory precursors are potentially affected by alcohol during the first trimester [46], these effects may be either largely negligible or may be corrected or compensated for to a large degree later during development. The observed hearing loss and prolonged ABR latencies following long-term PAE in rats [18‒21] and humans [3, 11‒13] are likely caused mostly by alcohol exposure during the second and/or third trimesters. Consistent with this idea, a single dose of alcohol at G12.5 resulted in CVG neuron loss [47].

ABR measures the synchronous electrical activity of auditory neurons and fibers; thus, the overall size, peak amplitudes of individual ABR waves, and the shape of ABRs can be used to access the integrity of the ascending auditory pathway. In this study, we measured the peak amplitude of ABR wave I because it can be most reliably identified across ages and groups. We found no significant effect of PAE on wave I amplitude at either P14 or 2 months of age, suggesting largely preserved organization of auditory pathways and neuron synchronization at least at the level of the auditory nerve. Importantly, we found that first-trimester PAE induced significant alterations in the size of ABRs, as evaluated by the absolute value of the integral of ABR waveforms, termed the ABR integral in this study. This measure describes the sum of the magnitudes of the waveform (not just the peak magnitudes), regardless of direction, providing a measure of the overall signal strength. At the age of 2 months, the ABR integral was significantly smaller in PAE offspring as compared to control offspring, and this effect occurred similarly in males and females (no effect of PAE × sex interaction). Although it does not probe specific auditory centers, the ABR integral may offer a more sensitive assessment of the overall integrity of the ascending auditory pathway at the level of the auditory nerve and the brainstem. PAE-induced ABR integral reduction may suggest a smaller number of active neurons in response to our click stimulation and/or compromised synchronization of active neurons. Either way, this observation demonstrates that first-trimester PAE induces delayed occurrence of auditory processing alterations in adults.

Comparing ABR components provided some clues about of potential cellular sites responsible for the altered ABR size. Although the reduction in ABR size was observed across waves, later waves III–V appeared to have more dramatic reductions. It is believed that waves III, IV, and V are generated from the superior olivary complex, the lateral lemniscus, and the inferior colliculus, respectively [48, 49], at least in typically developed mice. Altered ABR shape during the window of waves III–IV in the PAE group resembles the finding in a human infant study that PAE resulted in a lack of reproducibility or the absence of peak V. These observations from individuals and mice strongly implicate an impact of PAE at the level of superior olivary complex and lateral lemniscus. In further support, PAE resulted in less neurons in the medial superior olive in non-human primates [22]. Neurons in the nucleus laminaris, the avian analog of the mammalian medial superior olive, displayed shorter dendrites and abnormal evoked neuronal activity in chicken embryos [7]. Thus, the observed ABR changes may be induced by alterations in neuron numbers, neuron morphology, and/or connectivity in the superior olivary complex, which warrants future cellular analyses.

We did not detect a significant effect of first-trimester PAE on the size of acoustic startle, as evaluated by peak amplitude, at either P14 or 2 months of age, indicating no effect on this involuntary reflex reaction. PPI of the acoustic startle is considered a reflection of auditory gating function. We found no effect of first-trimester PAE on PPI triggered by 75- and 80-dB prepulses in mice, consistent with a report in which mice were treated with a binge-like PAE at G8 [50]. Disrupted PPI by 80-dB prepulses was reported in several studies of rhesus monkeys and mice with long-term PAE [25, 51], which may be a consequence of alcohol exposure during a later developmental stage.

At P14 but not at 2 months, we detected a significant effect of PAE treatment on PPI triggered by lower levels of prepulses. PAE offspring were less likely to display an effective PPI to 68- and 75-dB prepulses at P14, and this effect persisted until 2 months for the 68-dB prepulse. It is important to consider hearing ability when interpreting these results. PAE offspring had largely normal ABR threshold to clicks, indicating that they hear as well as control mice in a quiet environment such as that used for ABR tests. However, it is possible that PAE offspring have a compromised ability to differentiate two sound levels and thus a reduced ability to listen in noisy environments. This possibility is supported by abnormal speech perception in noise, a binaural processing deficit, in humans with PAE and is consistent with disorganized brainstem nuclei involved in binaural computation [5, 7, 22]. Because we tested ASRs and PPIs with a background of 65 dB noise (a commonly used standard), we cannot exclude the possibility that PAE-induced PPI disruptions may be attributed, partially or totally, to an auditory-related problem. Regardless of the relative contribution of hearing sensitivity with noise versus auditory gating function, our data demonstrate that first-trimester PAE induces auditory system and/or auditory processing dysfunction as early as the age of hearing onset and into adulthood.

We did not observe an effect of first-trimester PAE on the measures associated with anxiety or spontaneous activity, suggesting normal brain functions mediating these behaviors. This finding also supports the idea that the detected auditory system alterations are not derived from a higher level influence. We did detect fewer arm entries in the EZM in PAE offspring. The significance of this effect is not clear, given unchanged total time in open arms, but may be somewhat associated with exploration [52]. Additional behavioral tests targeting exploration would be needed to address this issue.

The results of this study should be considered with two limitations. First, the limited sample sizes may fail to detect some subtle changes. Indeed, some auditory-related measures show a trend of changes in the same direction as long-term PAE such as ABR thresholds. It is likely that some PAE effects are small and require a larger sample size to achieve significance. A good example of ABR size analysis supports this possibility. By analyzing individual ABR peak amplitudes, we could not detect a significant change. However, when analyzing the ABR integral that considered all alterations across waves, the PAE effect became statistically significant. The limited sample size also hindered our ability to determine whether a litter effect was present. Due to the small number of subjects, statistical power was insufficient to definitively assess the impact of litter variability on the observed outcomes. Consequently, further studies with larger sample sizes are necessary to accurately evaluate the potential confounding effect of litter and to obtain more robust conclusions regarding treatment effects. Second, the concentration of alcohol used in our PAE mice should be considered. Unfortunately, most studies investigating PAE effects on the auditory system in rodents did not provide BAC measures. The BAC level (119 mg/mL) measured in this study is considered a low to medium BAC (80–150 mg/mL) [53]. In humans, it is estimated that high BACs of over 200 mg/dL may be responsible for severe FAS phenotypes, while lower BACs may induce milder symptoms [54]. Additionally, we should consider the route of PAE when comparing across studies. A variety of methods have been used to achieve PAE in animal models including intraperitoneal injection, oral gavage, gastric catheter, and ethanol in drinking water, with some causing stress on pregnant dams and others causing chronic dehydration [53]. We used the Lieber-DeCarli diet to model PAE to reduce this potential confound, with the goal of enhancing the specificity of detected effects to actual alcohol exposure.

The translational implication of our study is considered in light of larger concerns about FASD and PAE and its relationship to socioeconomic status (SES). While prolonged alcohol consumption during pregnancy is widely known to result in numerous deficits, our data show that even moderate exposure during early pregnancy can significantly impair certain aspects of auditory system development. FASD ranks among the top three known causes of intellectual disability, with the current prevalence of FASD in the United States ranging from 2.4% to 4.8% [55]. Prior to the manifestation of measurable cognitive changes, auditory functional tests such as ABR and acoustic startle (called Moro reflex in infants), both of which are present at birth, have the potential to serve as early markers for diagnosing FASD and predicting its severity. Of note, PAE occurs across all socioeconomic strata, affecting both high and low SES groups. Although high SES groups appear to have a higher prevalence of alcohol consumption during pregnancy, there is a negative correlation between SES and the severity of FASD in the aspects of language development as well as memory and executive function [56], probably attributable to other factors associated with low SES, such as increased stress and undernutrition. A recent study reported that the effects of PAE may outweigh the impact of SES on brain subcortical volumes and neurocognitive abilities in typically developing individuals [57]. Given that early auditory experiences and processing are crucial for academic performance in school-aged children, PAE-induced auditory system deficits may present an important factor in understanding the relationship between PAE and the SES achievement gap and using early education and screening to help close that gap.

Summary: One of the most important tasks of hearing in humans is to enable communication through verbal speech, a task requiring both sufficient hearing sensitivity and accurate auditory processing in the central system. A study of school-aged children reported twice as many children with compromised central auditory processing as those showing reduced hearing sensitivity [58]. PAE affects both peripheral (likely attributable to prolonged exposures) and central (inducible by early exposure) auditory systems. We present evidence supporting an effect of first-trimester PAE on specific auditory processing properties that may involve neuron synchronization and/or listening in noise at the brainstem level. Such effects may begin as early as hearing onset and can last into young adulthood, supporting the potential of screening central auditory functions as early markers for FASD diagnosis, and hopefully, severity prediction.

We thank Dr. Terra Bradley (Florida State University) for the careful editing of the manuscript.

The study was conducted in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals and approved by the Florida State University Institutional Animal Care and Use Committee (IPRO-TO202200000025, approved on June 16, 2022).

The authors have no conflicts of interest to declare.

This research was funded by Florida State University’s Council on Research and Creativity Multidisciplinary Support Grant and Florida State University Rodgers and Rodgers Hearing Foundation.

Conceptualization: Jennifer Steiner and Yuan Wang; formal analysis: Mark Jessup, Maya Liu, Deirdre McCarthy, and Yuan Wang; funding acquisition: Jennifer Steiner and Yuan Wang; investigation: Mark Jessup, Abigail Tice, Addison McNeill, Avery Tangen, and Yuan Wang; writing – original draft: Mark Jessup and Yuan Wang; and writing – review and editing: Deirdre McCarthy, Pradeep Bhide, Jennifer Steiner, and Yuan Wang.

The main results of this study are reported within this manuscript. The original data that support the findings of this study are available from the corresponding author upon reasonable request.