Long-term effects of low prenatal alcohol exposure on GABAergic interneurons of the murine posterior parietal cortex

Abstract

Background:

Fetal Alcohol Spectrum Disorders (FASDs) are characterized by a wide range of physical, cognitive, and behavioral impairments that occur throughout the lifespan. Prenatal Alcohol Exposure (PAE) can lead to adult impairments in cognitive control behaviors known to be mediated by the posterior parietal cortex (PPC). The PPC plays a fundamental role during the performance of response tasks in both primates and rodents, specifically when choices between similar target and non-target stimuli are required. Furthermore, the PPC area is known to be reciprocally connected with other cortical areas. Despite the extensive literature investigating the molecular mechanisms underlying PAE impairments in cognitive functions mediated by cortical areas, little is known to date regarding the long-term effects of PAE on PPC development and function. Here, we examined changes in cellular organization of GABAergic interneurons and their function in PPC using behaviorally naïve control and PAE mice.

Methods:

We used a limited access model of PAE; C57BL/6J females were exposed to a solution of 10% (w/v) ethanol and 0.066% (w/V) saccharin for 4 hrs/day throughout gestation. Using high-throughput fluorescent microscopy, we quantified the levels of GABAergic interneurons in the PPC of adult PAE and control offspring. In a separate cohort, we recorded spontaneous inhibitory postsynaptic currents (sIPSCs) using whole-cell patch clamp recordings from PPC layer 5 pyramidal neurons.

Results:

We found that PAE led to a significant overall reduction of parvalbumin-expressing GABAergic interneurons in PAE mice regardless of sex. Somatostatin- and calretinin-expressing GABAergic interneurons were not affected. Interestingly, PAE did not modulate sIPSC amplitude and frequency.

Conclusions:

Together, these results suggest that impairments in cognitive control observed in FASD may be due to the significant reduction of parvalbumin-expressing GABAergic interneurons observed in PPC, and PAE animals may show compensatory changes in GABAergic function following developmental reduction of these interneurons.

Introduction

Fetal Alcohol Spectrum Disorders (FASDs) refers to a range of neurodevelopmental deficits caused by prenatal alcohol exposure (PAE) (Streissguth and O’Malley, 2000, Chudley et al., 2005). Despite the numerous recommendations to abstain, alcohol consumption during pregnancy is still common, with data collected in the United States from 2018 to 2020 showing that nearly 14% of pregnant women drink alcohol and ~5% report binge drinking in the past 30 days (Gosdin et al., 2022). As the National Survey on Drug Use and Health data collected in 2017-2018 showed that 89.3% of pregnant women reported abstinence, maternal drinking may be increasing (8 U.S. Department of Health and Human Services, Office of Disease Prevention and Health Promotion. Healthy People 2020: maternal, infant, and child health. MICH-11.1 Increase abstinence from alcohol among pregnant women. 2020).

Adolescents with FASDs commonly report difficulties in cognitive and executive functioning, including impairments in learning, working memory, response inhibition, and behavioral flexibility (Streissguth et al., 1991, Mattson et al., 1999, Green et al., 2009, Marquardt et al., 2021). Preclinical and clinical investigations performed to date show that exposure to alcohol during early development affects multiple brain areas that underlie these behaviors (Marquardt and Brigman, 2016). For example, cortical sub-regions involved in the regulation of cognitive and executive functions have consistently been shown to be affected by PAE (Friedman and Robbins, 2022, Miller and Cohen, 2001, Tanji and Hoshi, 2008, Jobson et al., 2021). Previously, we demonstrated that moderate PAE exposure (80-90 mg/dL) throughout gestation significantly impaired behavioral flexibility decreasing the orbitofrontal cortex (OFC) firing rate during the reversal learning phase (Marquardt et al., 2014, Marquardt et al., 2021). Later, Kenton and colleagues demonstrated a significant increase in calretinin-expressing cortical interneurons in the same brain area of adult male PAE offspring (Kenton et al., 2020), but no significant changes in the number of somatostatin or parvalbumin-expressing cortical interneurons in that region. They also found a significant increase in sIPSC amplitude and area in OFC pyramidal neurons (Kenton et al., 2020), suggesting that the reduction of glutamatergic activity in OFC observed during discrimination/reversal learning study (Marquardt et al., 2021) could be mediated by an increase in the function of GABAergic interneurons.

Recently, we showed that adult mice exposed to moderate levels of ethanol during development exhibited impaired cognitive control during the presentation of non-target trials on a touchscreen continuous performance task (5-Choice Continuous Performance Task (5C-CPT)(Olguin et al., 2021). Attention is a complex cognitive process that depends on differing cortical sub-regions, including the PPC. Lesions on this sub-region lead to a decrease in choice accuracy and an increase in perseverative responding in rats (Muir et al., 1996). It was recently shown that optimal performance of the 5C-CPT in mice also requires intake PPC function to inhibit the response to non-target stimuli (Young et al., 2020).

The PPC is commonly described as an association cortex, as it combines inputs from different brain areas including visual, somatosensory, auditory, visual and prefrontal cortices (Whitlock, 2017). Like all cortical areas, the PPC has a laminar organization and its functioning requires a balance between excitatory pyramidal neurons and inhibitory GABAergic interneurons. The majority of GABAergic interneurons expressed in cortex are parvalbumin-(PV⁺), calretinin-(CR⁺) and somatostatin-(SST⁺)-expressing interneurons (Xu et al., 2004, Butt et al., 2005, Miyoshi et al., 2007) and parvalbumin-expressing interneurons are the most highly expressed in the PPC (Perrenoud et al., 2013, Whissell et al., 2015, Hovde et al., 2019).

Despite the important role played by PPC in the cognitive attention processes, little is known about how PAE alters PPC structure and function during development. Therefore, we investigated the impact of a low PAE model on the number of PV, CR and SST-expressing interneurons in the PPC, as well as measuring the functional impact of developmental exposure using slice electrophysiological techniques.

Material and Methods

All experimental procedures were performed in accordance with the National Institutes of Health Guide for Care and Use of Laboratory Animals and were approved by the University of New Mexico Health Sciences Center Institutional Animal Care and Use Committee.

Prenatal Alcohol Exposure Paradigm

Control and PAE mice were generated by the New Mexico Alcohol Research Center, utilizing a limited-access drinking paradigm previously described (Brady et al., 2012). Briefly, female C57BL/6J mice (PND 60 days) (Jackson Laboratory, Bar Harbor, ME) were exposed to 0.066% (w/v) saccharin alone (SAC) (Sigma, product #S6047) or to Ethanol (EtOH) (KOPTEC, product #V1101) solution (5% w/v for 4 days, then 10% w/v) sweetened with 0.066% (w/v) saccharin for 4 hours per day (from 10:00 to 14:00 hours) during the dark cycle. Mice were housed under reverse 12 hr light/dark cycle (lights off 08:00 hr). Before and during the first 13 gestational days, pregnant females were on Teklad 2929 diet chow, while from gestational day 14 until the delivery mice received Tekland 7904 breeder chow.

After 7 days of acclimation to SAC or 10% EtOH solutions, individual females and singly housed male (Jackson Laboratory, Bar Harbor, ME) were bred together for 2 hours immediately following the drinking period (from 14:00 to 16:00 hours) for 5 consecutive days. Females continued the exposure throughout the 5-days mating period and were weighed every 3 to 4 days for determining pregnancy. Exposure continued throughout pregnancy and SAC and EtOH concentrations were halved every 2 days starting from postnatal day 1 (PND 1). Control and PAE offspring were weaned around PND 23 and housed in groups of 2 per cage. In the present study, the average of alcohol consumed by dams during the entire exposure was 5.37 ± 0.18 g/kg/4 hours and the blood ethanol concentration (BAC) measured with AM1 Alcohol Analyzer (Analox Instruments, USA) was 30.08 ± 3.19 mg/dl at the end of the 4 hours of drinking period. A cohort of 39 mice (9-10 animals for each experimental group) from 9 SAC and 9 PAE litters were used for immunohistochemically studies. These mice were food restricted to reach 85% of their body weight as previously described (Olguin et al., 2021), in order to reproduce the same experimental conditions. A separate cohort of 18 mice (4-5 animals for each experimental group) from 8 SAC and 8 PAE litters were used for electrophysiological studies. These mice had ad libitum access to food as previously described (Licheri et al., 2021) as food-restriction in mice or rats can alter cortical GABAergic transmission (Dazzi et al., 2014, Yang et al., 2016). In all cases, the investigators were blind to the treatment group assignments.

Brain slice preparation

Adult mice (age: 111.7 ± 3.12 days; weight SAC males: 32.45 ± 0.74, PAE males: 31.02 ± 0.75; weight SAC females: 22.13 ± 0.61, PAE females: 24.67 ± 0.32) were deeply anesthetized with ketamine (250 mg/kg intraperitoneal) and transcardially perfused with cold cutting solution containing in (mM): 92 N-methyl-D-glucamine (NMDG), 2.5 KCl, 10 MgSO₄, 0.5 CaCl₂, 1.25 NaH₂PO₄, 30 NaHCO₃, 20 Hepes, 25 Glucose, 2 Thiourea, 3 Na-Pyruvate, 5 Na-ascorbate (pH to 7.3-7.4 with HCl and bubbled with 95% O₂ and 5% CO₂, adjusted osmolarity to 290-310 mOsm) (Ting et al., 2018). The brain was rapidly removed from the skull and transferred for 2 min into ice-cold cutting solution. Coronal slices (300 μm thick) were cut with a Leica VT1000 plus vibratome (Leica Microsystems, Bannockburn, IL) and transferred to an incubation chamber held at 34 °C and containing NMDG cutting solution to which increasing amounts of NMDG-solution containing 2M NaCl were added in a step-wise fashion (250 μl at 0 min, 250 μl at 5 min, 500 μl at 10 min, 1 ml at 15 min, and 2 ml at 20 min) until a final concentration of 52 mM was reached. Slices were moved to recover in a holding chamber (model BSC-PC, Warner Instruments, Hamden, CT) with artificial cerebral spinal fluid (aCSF) containing in (mM): 92 NaCl, 2.5 KCl, 1.25 NaH₂PO₄, 30 NaHCO₃, 25 glucose, 20 HEPES, 2 Thiourea, 3 Na-Pyruvate, 5 Na-ascorbate, 2 CaCl₂, and 1 MgSO₄. (pH to 7.3-7.4 with a few drops of 1 M NaOH and bubbled with 95% O₂ and 5% CO₂, (290-310 mOsm). Slices were then kept at room temperature (approximately 22 °C) for a least 40-60 min before starting electrophysiological recordings.

Whole-cell recordings

After incubation, a hemi-slice was placed in a recording chamber with slice support (Cat # RC-27L; Warner Instruments) to allow flow of aCSF above and below the slice. The recordings were performed in the presence of aCSF containing in (mM): 126 NaCl, 3 KCl, 1.25 NaH2PO4, 1 MgSO₄, 2 CaCl₂, 26 NaHCO₃, 10 glucose, 5 Na-ascorbate, osmolarity 290-310 mOsm, pH 7.4 bubbled with 95% O₂ and 5% CO₂ and delivered to the chamber at a flow rate of 2 ml/min using a peristaltic pump (Master Flex, model 7518-10, Cole-Parmer, Vernon Hills, IL). Temperature was maintained constant at 34 °C with a dual automatic temperature controller (Model TC-344B) that was also connected to an in-line solution heater (Model SH-27B) (Warner instruments).

Cells located in layer V of PPC were identified using an Olympus BX50WI upright microscope (Olympus, Center Valley, PA) equipped with infra-red differential interference contrast optics connected to a charge-coupled device camera (CCD100, DAGE-MTI, Michigan City, IN). Recording pipettes were prepared from filament-containing borosilicate capillary glass (outside diameter, 1.5 mm, inside diameter 0.86 mm, catalog # BF150-86-10; Sutter Instruments) and pulled to a resistance between 2.5 and 5.0 MΩ using a DMZ Universal Electrode Puller (Zeitz-Instruments Vertriebs GmbH, Martinsried, Germany).

The internal pipette solution contained (in mM): 120 CsCl, 10 Hepes, 2 MgCl2, 1 EGTA, 2 MgATP, 0.3 mM NaGTP, 1 QX314, adjusted to 288 mOsm, pH 7.3-7.4 (Badanich et al., 2013). Excitatory and inhibitory PSCs were recorded in pyramidal neurons voltage-clamped at −70 mV; after the baseline stabilization, (typically 10 minutes after breaking into cell), the baseline activity was recorded for additional 5 minutes in presence of aCSF, then NMDA and AMPA receptor-mediated currents were blocked respectively with D-2-Amino-5-phosphonovaleric acid (D-APV 50 μM, Hello Bio) and 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX 10 μM, Hello Bio) and recorded for 10 additional minutes. Then, the GABA_A receptor antagonist picrotoxin (100 μM, Hello Bio) was added to block sIPSCs.

Whole-cell recordings were conducted with an Axopatch 200-B amplifier (Molecular Devices, Sunnyvale, CA). Currents were filtered at 2 kHz, and digitized at 5 kHz using an analog-to-digital signal converter Model 1440A (Molecular Devices, Sunnyvale, CA).The noise was reduced using a Hum Bug 50 Hz/60 Hz Noise Eliminator (Scientific Systems Design Inc, Ontario, Canada). Data were acquired using Clampex software version 10.7 (Molecular Devices, Sunnyvale, CA). Only recordings with a stable access resistance that varied < 25% were included in the analysis. Off-line analysis was performed using Minianalysis 6.0 (Synaptosoft). For statistical analysis, we included sIPSCs recorded during the last 3 min of the D-APV and CNQX bath-application. Cells recorded with capacitance values > 100 pF were considered pyramidal neurons (Kenton et al., 2020).

Separation of small and large events

To investigate if PAE specifically modulated large amplitude sIPSC events mediated by PV interneurons, an objective approach was used for calculating a threshold for large events based on the amplitude distribution of all events recorded from each cell (Chen et al., 2018). The cumulative distribution of all sIPSC amplitude was calculated in Minianalysis 6.0, then plotted and fitted with either one or two cumulative normal distribution using the distribution function in Excel. Next, the p-value was calculated using F-test Two-Sample for variances function in Excel, and a value close to 0 indicated that two cumulative amplitude distributions fit better the cumulative curves compared to one distribution. Finally, a threshold of amplitude, for separating small and large events, was identified using mathematical analysis performed to calculate the x value (representing amplitude) corresponding to the y value (representing cumulative distribution) where the first distribution ends.

Triple staining immunohistochemistry

As mentioned above, adult control and PAE mice were food restricted at 7 weeks after birth. Males were food-restricted for 14 days, while female groups took ~24 days (age: male: 66 ± 0.66 days; female: 98.50 ± 1.80 days; weight SAC males: 21.97 ± 0.31, PAE males: 20.78 ± 0.41; weight SAC females: 20.07 ± 0.49, PAE females: 19.55 ± 0.21). Mice were deeply anesthetized with ketamine (250 mg/kg intraperitoneal) and intra-cardiac perfused with 4% w/v paraformaldehyde (PFA (J.T. Baker Chemical Co.), in phosphate buffered saline (PBS) (Thermo-Fischer Scientific, USA), pH 7.4. After extraction, brains were stored in PFA for 48 hours at 4 °C and then in 1X PBS. Coronal slices containing PPC were cut at 50 μm in thickness using a vibratome (PELCO easiSlicer, Ted Pella, INC. Redding, CA, USA) and stored in 24-well plates containing 1X PBS. During day 1 floating sections were incubated for 1.5 hour with 1% bovine serum albumin (Sigma Aldrich), 5% donkey serum (Jackson ImmunoResearch, USA), 0.2% Triton X-100 (Sigma Aldrich) in PBS (pH 7.4). Next, the sections were incubated with 300 μl of primary antibody solution: 1:5000 dilution of anti-parvalbumin mouse monoclonal antibody (catalog # 235, Swant, Bellinzona, Switzerland.), 1:1000 dilution of anti-somatostatin rabbit polyclonal antibody (catalog # , T-4103BMA Biomedicals, Switzerland), 1:1000 dilution of anti-calretinin chicken polyclonal antibody (catalog # CH22116, Neuromics, Edina, Minnesota, USA) for 7 nights at 4 °C. On the 8^thday, sections were washed 4 times with 1X PBS and incubated for 30 min with blocking solution without 0.2 % Triton X-100. After blocking, 1:1000 dilution of Alexa Fluor 488 (donkey anti-mouse IgG, Thermo Fischer Scientific, USA), 1:1000 dilution of Alexa Fluor 555 (donkey anti-rabbit IgG, Thermo-Fisher Scientific, USA) and 1:1000 dilution of Alexa Fluor 647 (donkey anti-chicken IgG, Jackson ImmunoResearch, USA) in blocking solution were added for 1.5 hours at room temperature. Sections were then washed 4 times with 1X PBS and incubated for 20 min with 4’,6-diamidino-2-phenylindole (DAPI, dilution 1:500, Thermo-Fischer Scientific, USA). Finally, sections were rinsed with 1X PBS and mounted on Superfrost Plus microscope slide (VWR, Radnor, PA) using Fluoromount-G mounting media (Thermo Fischer Scientific, USA), covered with glass coverslips (VWR, Radnor, PA) and stored at 4 °C until imaging acquisition.

Coronal sections were scanned by an Axio Scan.Z1 slide scanner (Zeiss, Germany) using a 20X objective (1X magnification), with the following filter settings: Alexa Fluor 405–LED 385 nm; Alexa Fluor 488–LED 475 nm; Alexa Fluor 555–LED 555 nm and Alexa Fluor 647–LED 630 nm. The counting of GABAergic interneurons was performed using the QuPath’s Cell Detection algorithm (Version 0.4.3), separating the superficial (layers II/III) from the deep cortical layers IV/VI (Fig. 1A) to characterize possible PAE effects on the migratory pathways during the development of the cortex.

Figure 1. Triple Staining Immunohistochemistry of GABAergic interneurons in PPC layer II/III and IV/VI after PAE.

A. Coronal section of adult brain stained with DAPI showing the location of PPC and its boundaries with other cortical regions and hippocampus. White-red dashed lines indicate the borders with the retrosplenial cortex (right) and visual cortex (left). White dashed line separates superficial layers (II/III) from deep layers (IV/VI). Scale bar is 400 μm. B-C. Representative image of coronal slices stained with DAPI from SAC (upper) and PAE (lower) coronal slices. D-E. Immunostaining for Parvalbumin (PV⁺) expressing GABAergic interneurons (green) shows the overall decrease in PAE (lower) mice across all cortical layers. F-G. Images comparing SAC (upper) and PAE (lower) coronal sections labeled for Calretinin (CR⁺) expressing GABAergic interneurons (red). H-I. Somatostatin (SST⁺) GABAergic interneurons in SAC (upper) and PAE (lower) slices. N= 8-10 mice for each experimental group with 1-2 sections per brain. Scale bar for all images is 200 μm.

For each class of interneurons analysed, the optimal cell detection parameters were set using the intensity threshold, while the other parameters were left as their default levels. The total number of GABAergic interneurons was calculated by adding the counts obtained from each section. The sections used for the counting were selected according to The Mouse Brain in Stereotaxic Coordinates (Paxinos, 2001) relative to bregma (AP: −1.94, ML 1.50). For each animal, 2 coronal sections were chosen and PPC was demarcated as lying between retrosplenial cortex and visual cortex and above the oriens-layer of the hippocampus (Fig. 1).

Drugs

All reagents and drugs used were purchased from Sigma (Sigma-Aldrich) unless otherwise indicated and most were dissolved in MilliQ water for making stock solutions since they were hydrochloride salts. Picrotoxin was dissolved in methyl sulfoxide (DMSO) (Thermo-Fisher) > 99.9% as stock solution, and after dilution, DMSO concentration was less than 0.1 %. All stock solutions were frozen in aliquots and before each single recording were diluted in aCSF to their final concentrations.

Statistical analysis

Data are expressed as means ± SEM. Statistical analyses for electrophysiological recordings were performed using Prism version 9.5.1 (GraphPad Software, San Diego, CA) and SPSS version 28 (IBM, Armonk, NY). Immunohistochemistry data were analysed using two-way analysis of variance (ANOVA) in Prism following meeting requirements for normality via the Shapiro-Wilk test. Possible outliers were detected using Rout test 1% and 5 outliers were removed from electrophysiological recordings, while we did not find any outliers in the immunohistochemistry data.

The sIPSCs data were analysed using a Linear Mixed Model in SPSS (Golub and Sobin, 2020). For each electrophysiological parameter analysed we built a specific model in a step-wise fashion using method adopted from Linear Mixed Models: A Practical Guide Using Statistical Software (West et al., 2007). First, we built a LMM including the animal random effect associated with intercept and using homogeneous residual error variances between experimental groups. Next, we performed a LMM excluding the random effect, and then using the −2-log restricted maximum likelihood value obtained from each model we performed a likelihood ratio chi-square test. If the p-value obtained was < 0.05, the random effect was included in the final model. Alternatively, if the p-value was > 0.05, the random effect was excluded from the final model. Next, it was necessary to test the appropriateness of inclusion or exclusion of heterogeneous residual error variances for each experimental group. For this purpose, another likelihood ratio chi-square test was performed and the heterogeneous residual error variances were included or excluded if the p-value was < 0.05 or > 0.05, respectively. Tables 1, 2 and 3 report F ratios (calculated with Satterthwaite approximated degrees of freedom) and p-values for Type III F-test for exposure, sex and exposure interaction obtained performing the final LMM. The effect sizes for exposure and sex were calculated using Hedges’s g formula and reported in Tables 1, 2 and 3 as described in our previous studies (Bird et al., 2021, Licheri et al., 2021). In addition, a Shapiro-Wilkes test was used for checking the normality of residuals and in case of p-value < 0.05 (see Tables 1, 2 and 3), also we performed non-parametric Mann-Whitney U test for fixed effects including effect size as $r=\frac{Z}{\surd N}$, where Z is Standard statistic, and N is the total of measurements (Tables 1, 2 and 3).

Table 1. Linear Mixed-Model Analysis of sIPSCs.

Table indicating for each variables analyzed if random effect of cells or heterogeneous error variances were used for building the model used for the statistical analysis. F-ratios and p-values are reported for Sex, Exposure and Sex*Exposure interaction effects. Hedge’s g effect size included for Sex and Exposure effects. Mann-Whitney U tests (r indicating effect size) are showed for Sex and Exposure effects when residuals violate the assumption of normality assessed using Shapiro-Wilk test.

Dependent Variable	Random effect of animal	Heterogeneous or Homogeneous	Exposure	Sex	Sex*Exposure	SW normality test	Mann-Whitney U: Exposure	Mann-Whitney U: Sex
Amplitude small events	Excluded	Heterogeneous	F(1,31)=0.525 p=0.474 g=0.218	F(1,31)=0.234 p=0.632 g=0.07	F(1,31)=0.121 p=0.731	Violated	U(n1=11,n2=24)=128 p=0.903 r=0.024	U(n1=20,n2=15)=123 p=0.382 r=0.016
Rise Time small events	Included	Homogeneous	F(1,18.013)=0.026 p=0.873 g=0.224	F(1,18.013)=0.787 p=0.387 g=0.452	F(1,18.013)=0.018 p=0.895	Violated	U(n1=10,n2=24)=115.5 p=0.867 r=0.029	U(n1=20,n2=14)=102 p=0.192 r=0.228
Decay Time small events	Excluded	Homogeneous	F(1,31)=0.751 p=0.393 g=0.432	F(1,31)=1.068 p=0.309 g=0.535	F(1,31)=0.442 p=0.511	Violated	U(n1=11,n2=24)=112 p=0.494 r=0.120	U(n1=20,n2=15)=108 p=0.169 r=0.236
Frequency small events	Excluded	Heterogeneous	F(1,31)=0.007 p=0.934 g=0.152	F(1,31)=8.842 p=0.006 g=0.757	F(1,31)=0.580 p=0.452	Violated	U(n1=11,n2=24)=107 p=0.390 r=0.150	U(n1=20,n2=15)=89 p=0.043 r=0.343
Charge small events	Excluded	Heterogeneous	F(1,30)=0.972 p=0.332 g=0.307	F(1,30)=0.196 p=0.661 g=0.047	F(1,31)=0.175 p=0.678	Violated	U(n1=11,n2=23)=121 p=0.856 r=0.034	U(n1=19,n2=15)=137 p=0.864 r=0.032

Table 2. Linear Mixed-Model Analysis used for comparing small sIPSCs between SAC and PAE groups.

Statistical values are reporting for each variable analyzed during the building of Linear Mixed-Models.

Dependent Variable	Random effect of animal	Heterogeneous or Homogeneous	Exposure	Sex	Sex*Exposure	SW normality test	Mann-Whitney U: Exposure	Mann-Whitney U: Sex
Amplitude large events	Excluded	Homogeneous	F(1,31)=0.931 p=0.342 g=0.410	F(1,31)=0.011 p=0.916 g=0.165	F(1,31)=1.379 p=0.249	Violated	U(n1=11,n2=24)=107 p=0.390 r=0.150	U(n1=20,n2=15)=146 p=0.908 r=0.022
Rise Time large events	Excluded	Homogeneous	F(1,31)=1.739 p=0.197 g=0.389	F(1,31)=0.787 p=0.387 g=0.071	F(1,18.013)=0.088 p=0.769	Not violated
Decay Time large events	Excluded	Homogeneous	F(1,31)=2.432 p=0.129 g=0.658	F(1,31)=0.351 p=0.558 g=0.475	F(1,31)=2.043 p=0.163	Not violated
Frequency large events	Excluded	Heterogeneous	F(1,11)=0.557 p=0.471 g=0.364	F(1,11)=2.355 p=0.153 g=0.643	F(1,11)=0.324 p=0.581	Not violated
Charge large events	Excluded	Homogeneous	F(1,31)=0.782 p=0.383 g=0.408	F(1,31)=0.170 p=0.683 g=0.370	F(1,31)=1.686 p=0.204	Violated	U(n1=11,n2=24)=108 p=0.409 r=0.144	U(n1=20,n2=15)=134 p=0.610 r=0.09

Table 3. Linear Mixed-Model Analysis used for comparing large sIPSCs between SAC and PAE groups.

Statistical values are reported for each variable analyzed during the building of Linear Mixed-Models.

Dependent Variable	Random effect of animal	Heterogeneous or Homogeneous	Exposure	Sex	Sex*Exposure	SW normality test	Mann-Whitney U: Exposure	Mann-Whitney U: Sex
Capacitance	Excluded	Homogeneous	F(1,31)=0.097 p=0.757 g=0.08	F(1,31)=0.294, p=0.591 g=0.12	F(1,31)=0.200 p=0.658	Violated	U(n1=11,n2=24)=116 p=0.587 r=0.096	U(n1=20,n2=15)=123 p=0.382 r=0.15
Membrane Resistance	Excluded	Homogeneous	F(1,31)=0.546 p=0.466 g=0.004	F(1,31)=0.010, p=0.919 g=0.283	F(1,31)=0.002 p=0.962	Not Violated
Amplitude	Excluded	Homogeneous	F(1,31)=1.544 p=0.223 g=0.514	F(1,31)=0.010, p=0.920 g=0.132	F(1,31)=0.672 p=0.418	Not Violated
Rise Time	Included	Homogeneous	F(1,17.509)=0.058 p=0.812 g=0.029	F(1,17.509)=0.462, p=0.506 g=0.334	F(1,17.509)=0.282 p=0.602	Violated	U(n1=11,n2=24)=119 p=0.662 r=0.078	U(n1=20,n2=15)=113 p=0.227 r=0.208
Decay Time	Included	Homogeneous	F(1,18.074)=0.739 p=0.401 g=0.091	F(1,18.074)=0.495 p=0.491 g=0.635	F(1,18.074)=0.407 p=0.531	Violated	U(n1=11,n2=24)=96 p=0.211 r=0.216	U(n1=20,n2=15)=110 p=0.191 r=0.225
Frequency	Excluded	Heterogeneous	F(1,31)=0.017 p=0.896 g=0.164	F(1,31)=8.594 p=0.006 g=0.756	F(1,31)=0.728 p=0.400	Violated	U(n1=11,n2=24)=109 p=0.430 r=0.138	U(n1=20,n2=15)=92 p=0.055 r=0.326
Charge	Excluded	Heterogeneous	F(1,30)=3.502 p=0.071 g=0.658	F(1,30)=0.188 p=0.667 g=0.324	F(1,30)=1.686 p=0.204	Violated	U(n1=11,n2=23)=75 p=0.06 r=0.325	U(n1=19,n2=15)=113 p=0.319 r=0.175

Results

PAE does not affect interneuron populations in the layers II/III of PPC

To investigate the effects of PAE on PPC layers containing cortical and thalamic inputs, we first counted cells in layer II/III. A two-way ANOVA analysis showed no significant main effects of exposure, sex or sex*exposure interaction for DAPI positive nuclei counted in the layers II/III (exposure effect: F(1,32) = 1.901, p-value = 0.1775; sex effect: F(1,32) = 0.4820, p-value = 0.4925; sex*exposure effect: F(1,32) = 0.04043, p-value = 0.8419) (Fig. 2A).

Figure 2. PAE induces a significant overall reduction of PV+-expressing interneurons in PPC.

A. Bar graphs showing the mean of total DAPI⁺ cells counted in PPC layer II/III. B. Comparison of PV⁺-expressing GABAergic interneurons between all experimental groups counted in the superficial layers of PPC. C. Mean of total number of CR⁺-expressing GABAergic interneurons expressed in layer II/III. D. Bar plot reporting the mean of total number of SST⁺-expressing GABAergic interneurons counted in superficial layers of PPC. E. Total number of DAPI⁺ cells counted in layer IV/VI of PPC. F. PAE does not significantly decrease the number of PV⁺-expressing GABAergic interneurons in the deepest PPC layers. G. Bar graphs report the mean number of CR⁺-expressing GABAergic interneurons counted in PPC layer IV/VI. H. No significant difference between SAC and PAE in the number of SST⁺-expressing GABAergic interneurons in the deepest layers of the brain area examined. Data are expressed as mean ± SEM. I. Bar graphs showing the mean of total DAPI-positive nuclei counted across all layers. J. PAE significantly decreases the global number of PV⁺-expressing GABAergic interneurons in PPC of adult offspring. K. The mean of total CR⁺-expressing GABAergic interneurons counted in all PPC layers. L. SAC and PAE groups do not show significant difference in the mean of total SST⁺-expressing GABAergic interneurons in whole PPC. Data are expressed as mean ± SEM. E indicates the main effect of Exposure. * p value < 0.05

Additionally, we did not observe any significant difference in the number of PV⁺ interneurons (exposure effect: F(1,32) = 3.832, p-value = 0.0590; sex effect; F(1,32) = 0.2379, p-value = 0.6290; sex*exposure effect: F(1,32) = 0.5301, p-value = 0.4719) (Fig. 2B). The statistical analysis performed for CR⁺ interneurons did not reveal any significant alteration (exposure effect: F(1,32) = 0.07011, p-value = 0.7929; sex effect: F(1,32) = 1.030, p-value = 0.3178; sex*exposure effect: F(1,32) = 0.5266, p-value = 0.4733) (Fig. 2C). Furthermore, the counting of SST⁺ interneurons expressed in PPC superficial layers did not show any significant change (exposure effect: F(1,32) = 0.04162, p-value = 0.8396; sex effect: F(1,32) = 1.382, p-value = 0.2485; sex*exposure effect: F(1,32) = 0.4825, p-value = 0.4923) (Fig. 2D).

PAE does not affect interneuron populations in the layers IV/VI of PPC

Next, we analyzed layer IV/VI of PPC. Analysis of DAPI staining did not show any significant alteration (exposure effect: F(1,33) = 1.133, p-value = 0.2948; sex effect; F(1,33) = 0.06570, p-value = 0.7993; sex*exposure: F(1,33) 0.1960, p-value = 0.6609) (Fig. 2E). Also, we did not observe significant changes in the number of PV⁺ interneurons expressed in the deepest layers of PPC (exposure effect: F(1,33) = 3.679, p-value = 0.0638; sex effect: F(1,33) = 0.3076, p-value = 0.5829; sex*exposure effect: F(1,33) = 0.4488, p- value = 0.5076) (Fig. 2F).

The PAE paradigm did not affect also the number of CR⁺interneurons (exposure effect: F(1,33) = 0.01310, p-value = 0.9096; sex effect: F(1,33) = 0.001649, p-value = 0.9679; sex*exposure effect: F(1,33) = 0.01146, p-value = 0.9154) (Fig. 2G). Additionally, we did not find any significant alteration in SST⁺ interneurons (exposure effect: F(1,33) = 0.3230, p-value = 0.5736; sex effect: F(1,33) = 2.372, p-value = 0.1331; sex*exposure effect: F(1,33) = 0.7080, p-value = 0.4062) (Fig. 2H).

PAE decreases the global number of PV⁺ interneurons in PPC

Although we did not observe significant changes in the GABAergic populations expressed in the superficial and deepest cortical layers of PPC, we counted the total number of GABAergic interneurons expressed in this brain area. The two-way ANOVA did not reveal significant main effects in the total number of DAPI positive nuclei counted in all layers (exposure effect: F(1,33) = 2.331, p-value = 0.1364; sex effect: F(1,33) = 0.5704, p-value = 0.4555; sex*exposure effect: F(1,33) = 0.001492, p-value = 0.9694) (Fig. 2I). Interestingly, PAE induced a significant global decrease of PV⁺ interneurons (exposure effect: F(1,33) = 4.328, p-value = 0.0453; sex effect: F(1,33) = 0.06248, p-value = 0.8042; sex*exposure effect: F(1,33) = 0.3986, p-value = 0.5322) (Fig. 2J). While, the PAE did not affect the total number of CR⁺ interneurons expressed in PPC (exposure effect: F(1,33) = 0.002832, p-value = 0.9579; sex effect: F(1,33) = 1.022, p-value = 0.3193; sex*exposure effect: F(1,33) = 0.7032, p-value = 0.4077) (Fig. 2K). Also, the total number of SST⁺ interneurons did not show any significant change (exposure effect: F(1,32) = 0.2129, p-value = 0.6476; sex effect: F(1,32) = 2.703, p-value = 0.1099; sex*exposure effect: F(1,32) = 0.2129, p-value = 0.6476) (Fig. 2L).

PAE does not modulate GABAA receptor-mediated sIPSCs

To investigate whether the PAE model modulates GABAergic activity into adulthood, we measured sIPSCs from pyramidal neurons located in layer V of PPC. The linear mixed model analysis did not show any significant alteration in the capacitance of PAE group compared to the control group (LMM: exposure effect: F(1,31) = 0.097, p-value = 0.757, Hedges’s g = 0.08; sex effect: F(1,31) = 0.294, p-value = 0.591, Hedges’s g = 0.12; sex*exposure effect: F(1,31) = 0.200, p-value = 0.658 (Fig. 3A). Including the random effect of animal did not significantly improve LMM and was not included in the final model. Similarly, the likelihood ratio test indicated the random effect of animal did not improve the analysis of membrane resistance (p-value was >0.05). The final model revealed that PAE did not affect this parameter in all experimental groups (LMM: exposure effect: F(1,31) = 0.546, p-value = 0.4666, Hedges’s g = 0.004; sex effect: F(1,31) = 0.010, p-value = 0.919, Hedges’s g = 0.283; sex*exposure effect: F (1,31) = 0.002, p-value = 0.962) (Fig. 3B). LMM analysis of sIPSC amplitude did not find significant differences for each fixed effect (LMM: exposure effect: F(1,31) = 1.544, p-value = 0.223, Hedges’s g = 0.514; sex effect: F(1,31) = 0.010, p-value = 0.920, Hedges’s g = 0.132; sex*exposure effect: F(1,31) = 0.672, p-value = 0.418) (Fig. 3C). Again, the final LMM excluded the random effect of animal (p-value > 0.05). LMM analysis of sIPSC frequency showed a significant sex effect (LMM: exposure effect: F(1,31) = 0.017, p-value = 0.896, Hedges’s g = 0.164; sex effect F(1,31) = 8.594, p-value = 0.006, Hedges’s g = 0.756; sex*exposure effect F(1,31) = 0.728, p-value = 0.400) (Fig. 3D). The likelihood ratio test performed for sIPSC rise (including random effect of animal) again did not detect significant statistical difference in all experimental groups (LMM: exposure effect: F(1,17.509) = 0.058, p-value = 0.812, Hedges’s g = 0.029; sex effect: F(1,17.509) = 0.462, p-value = 0.506, Hedges’s g = 0.334; sex*exposure effect: F(1,17.509) = 0.282, p-value = 0.602) (Fig. 3E). The sIPSC decay time (including the random effect of animal) revealed no significant change between PAE and control groups (LMM: exposure effect: F(1,18.074) = 0.739, p-value = 0.401, Hedges’s g = 0.550; sex effect: F(1,18.074), p-value = 0.491, Hedges’s g = 0.635; sex*exposure F(1,18.074) = 0.407, p-value = 0.531) (Fig. 3F).

Figure 3. Effects of prenatal alcohol exposure on spontaneous inhibitory postsynaptic currents (sIPSCs) recorded from pyramidal neurons located in layer V of PPC.

A. The first and second-trimester equivalent alcohol exposure does not alter the capacitance of PPC pyramidal neurons recorded in adult offspring. B. Membrane resistance of PPC pyramidal neurons does not change in PAE adult mice compared to SAC. C. Low PAE does not affect sIPSC amplitude. D. Bar graphs reporting the Sex effect on sIPSC frequency. E and F. Rise and decay-time measured in pyramidal neurons from slices coming from all experimental groups. G. Comparison of sIPSC charge transfer between SAC and PAE groups. H. Representative traces showing the sIPSCs recorded from PPC pyramidal neurons in SAC (black) and PAE (red) groups. Traces at the top show sIPSCs from male groups. Traces at the bottom show sIPSCs from female groups. Scale bar: 200 pA, 5 s. The number of recordings is indicated with black dots for males and white dots for females. Data are expressed as mean ± SEM. S indicates the main effect of Sex. * p value < 0.05.

As reported for sIPSC amplitude and kinetic parameters, PAE did not lead to significant statistical difference on sIPSC charge in all experimental groups (LMM: exposure effect F(1,30) = 3.502, p-value = 0.071, Hedges’s g = 0.658; sex effect F(1,30) = 0.188, p-value = 0.667, Hedges’s g = 0.324; sex*exposure effect F(1,30)= 1.686, p-value = 0.204) (Fig. 3G).

PAE does not alter large sIPSC mediated by PV⁺ interneurons

It is well established in literature that PV⁺ interneurons are involved in the generation of large sIPSCs, and for this purpose we investigated whether PAE would differentially affect small and large sIPSCs. As mentioned in the Material and Methods section, we separated sIPSCs into two subgroups in relation to their amplitudes. Analysis of sIPSC amplitude of small events (excluding random effects) did not reveal significant change in PAE group compared to control group (LMM: exposure effect F(1,30) = 0.972, p-value = 0.332, Hedges’s g = 0.218; sex effect F(1.30) = 0.196, p-value = 0.661, Hedges’s g = 0.072; sex*exposure effect F(1,30) = 0.175, p-value = 0.678) (Fig. 4A). In addition, the LMM analysis of the sIPSC amplitude of large events did not show significant alteration in each experimental group (LMM: exposure effect: F(1,31) = 0.931, p-value = 0.342, Hedges’s g = 0.410; sex effect: F(1,31) = 0.011, p-value = 0.916, Hedges’s g = 0.165; sex*exposure effect: F(1,31) = 1.379, p-value = 0.249) (Fig. 4F). Interestingly, the statistical analysis of sIPSC frequency (random effects excluded) in the subgroup of small events revealed a significant sex effect (LMM: exposure effect: F(1,31) = 0.007, p-value = 0.934, Hedges’s g = 0.152; sex effect: F(1,31) = 8.842, p-value = 0.006, Hedges’s g = 0.757; sex*exposure effect F(1,31) = 0.580, p-value = 0.452) (Fig. 4B). Also, LMM for sIPSC frequency in the subgroup of large events (excluding random effect of animal) did not reveal significant difference in each experimental group (LMM: exposure effect: F(1,11) = 0.557, p-value = 0.471, Hedges’s g = 0.364; sex effect: F(1,11) = 2.355, p-value = 0.153, Hedges’s g = 0.643; sex*exposure effect: F(1,11) = 0.324, p-value = 0.581) (Fig. 4G).

Figure 4. Small and large spontaneous inhibitory postsynaptic currents on PPC layer V pyramidal neurons are not affected by prenatal alcohol exposure.

A. Prenatal alcohol exposure does not modulate sIPSC amplitude of small events. B. Bar graphs showing the significant Sex effect on sIPSC frequency. C. Charge of small sIPSC results not to be modulated by prenatal alcohol exposure. D and E. Prenatal alcohol exposure does not modulate the kinetic parameters of small events mediated by GABAA receptors in PPC pyramidal neurons. F. sIPSC amplitude of large events is not affected by prenatal alcohol exposure. G. Frequency mediated by large sIPSCs does not show exposure effect between SAC and PAE groups. H. Bar graphs reporting no significant alterations on charge of large sIPSCs. I and J. Rise and decay-time of large sIPSCs are not affected by prenatal alcohol exposure. K. Example showing the objective analysis used to divide sIPSCs in small and large groups. The cumulative distribution of sIPSC amplitude was fitted with one (dashed line) and two (solid line) cumulative distributions. The amplitude threshold was calculated using the value on the x-axis on the fitted curve corresponding to the start of the second distribution. L. Representative traces showing small events of sIPSCs in all experimental groups. M. Representative traces showing large events of sIPSCs in SAC and PAE groups. Scale bar: 50pA, 20 ms. The number of recordings is indicated with black dots for males and white dots for females. Data are expressed as mean ± SEM. S indicates the main effect of Sex. * p value < 0.05.

The charge of small sIPSC events was not affected by gestational alcohol exposure. The statistical analysis performed (excluding the random effect of animal) did not reveal significant difference between each experimental group (LMM: exposure effect: F(1,30) = 0.972, p-value = 0.332, Hedges’s g = 0.307; sex effect: F(1,30) = 0.196, p-value = 0.661, Hedges’s g = 0.047; sex*exposure F(1,30) = 0.175, p-value = 0.678) (Fig. 4C). Also, the charge of large events (excluding the random effects) did not show significant alteration in PAE vs control groups (LMM: exposure effect: F(1,31) = 0.782, p-value = 0.383, Hedges’s g = 0.408; sex effect F(1,31) = 0.170, p-value = 0.683, Hedges’s g = 0.370; sex*exposure effect: F(1,31) = 1.686, p-value = 0.204) (Fig. 4H).

The likelihood ratio test calculated for building the LMM for sIPSC rise time of small events (including the random effect of animal) showed that PAE did not alter this kinetic parameter (LMM: exposure effect: F(1, 18.013) = 0.026, p-value = 0.873, Hedges’s g = 0.224; sex effect: F(1, 18.103) = 0.787, p-value = 0.387, Hedges’s g = 0.452; sex*exposure effect: F(1, 18.013) = 0.018, p-value = 0.895) (Fig. 4D). The random effect of animal was excluded from the final LMM for sIPSC rise time of large event, and did not indicated significant difference between PAE and control groups (LMM: exposure effect: F(1,31) = 1.739, p-value = 0.197, Hedges’s g = 0.389; sex effect: F(1,31) = 0.088, p-value = 0.769, Hedges’s g = 0.071; sex*exposure effect: F(1,31) = 4.054, p-value = 0.053) (Fig. 4I).

The models built for investigating if PAE altered the decay time of small and large events did not reveal significant difference between groups. In-detail analysis for small events (LMM for small: exposure effect: F(1,31) = 0.751, p-value: 0.393, Hedges’s g = 0.432; sex effect: F(1,31) = 1.068, p-value = 0.309, Hedges’s g = 0.535; sex*exposure: F(1,31) = 0.442, p-value = 0.511. LMM for large: exposure effect: F(1,31) = 2.432, p-value = 0.129, Hedges’s 0.658; sex effect: F(1,31) = 0.351, p-value = 0.558, Hedges’s g = 0.475; sex*exposure: F(1,31) = 2.043, p-value = 0.163 (Fig. 4E and J)). In both final models, the random effect was excluded considering that the p-value calculated through likelihood ratio test was > 0.05.

Discussion

In the current study, triple immunohistochemistry performed on coronal sections of 3 months-old offspring revealed that low PAE during first- and second- trimester equivalents induced a significant overall reduction in the number of PV⁺ GABAergic interneurons in PPC without any significant alteration in the total number of cells labeled with DAPI. The reduction in PV⁺ GABAergic interneurons was not specific to layer II/III or IV/VI, but was generally decreased across all layers, leading to a significant overall reduction in both male and female PAE mice. To our knowledge, the present investigation is the first evaluating the long-term effects of voluntary PAE on GABAergic populations of murine PPC. To date, there is only one preclinical study performed in rats showing that alcohol exposure during third- trimester equivalent did not affect the total number of neurons and glial cells in PPC measured in slices coming from PND 10 pups (Mooney and Napper, 2005). Despite the different animal model and paradigm used, together these results suggest that gestational alcohol exposure selectively affects specific cell types of this cortical sub-region. Furthermore, the reduction of PV⁺ GABAergic interneurons observed in our experimental model adds weight to the existing literature showing that altered excitatory and inhibitory (E/I) balance within cortical circuit is involved in the behavioral and cognitive deficits observed in FASD models (Sadrian et al., 2013).

Our current data revealed a selective and specific effect on PV⁺ GABAergic interneurons in the PPC, with no significant effect of PAE on SST⁺ or CR⁺ GABAergic populations globally in PPC or in specific layer. Previous studies have shown that early postnatal alcohol exposure during PND 7 (2.5 g/kg, two doses, two hours apart) caused a 30% reduction of cortical CR⁺ and PV⁺ GABAergic cortical interneurons in adult mice (Smiley et al., 2015). While PND 5, 7, and 9 alcohol exposure (5 g/kg, two doses, two hours apart) significantly decreased the number of mPFC PV⁺ GABAergic interneurons measured at PND 83 (Hamilton et al., 2017). The pattern of interneuronopathy seen in the current study, i.e. non-significant reductions in layers II/III and IV/VI, which combine into a global reduction in PV+ interneurons in PPC, suggests that low PAE can impair cortical development both in neurogenesis and integration, but not migration. Previous studies utilizing gestational age-delimited exposure to alcohol found significant overall enhancement in the total number of PV⁺ GABAergic interneurons in layer V of mPFC due to mis-migration (Skorput et al., 2015, Skorput and Yeh, 2016). Similarly, our previous investigation found that limited-access moderate PAE significantly increased the number of CR⁺, but not significant changes for PV⁺ and SST⁺ GABAergic interneurons in the orbitofrontal cortex (Kenton et al., 2020).

Together, these findings suggest that cortical GABAergic interneurons are selectively sensitive to alcohol exposures during specific phases of gestation that can lead to differential effects on migration and survival. Previous work had shown that alcohol displays its toxicity on neurogenesis and migration, processes that occurs from embryonic day 10 to 17 (Miyoshi and Fishell, 2011), while postnatal exposure leading to cell loss considering that cortical neurons and interneurons have reached their final position in all cortical layers (Rymar and Sadikot, 2007, Miyoshi et al., 2010). Our finding adds weight to the existing literature showing that FASD deficits depend on several factors such as exposure, timing, dose, frequency of alcohol exposure and route of administration (Kobor and Weinberg, 2011).

In a separate cohort of behaviorally naïve PAE animals, we examined whether the overall reduction of PV⁺ GABAergic interneurons altered GABAergic transmission in PPC. Whole-cell patch recordings obtained from the pyramidal neurons of PPC layer V did not reveal any significant alterations in sIPSCs measured in PAE adult offspring (PND ~111), but only a significant sex effect on sIPSC frequency. As PV⁺ GABAergic interneurons produce strong inhibitory synapses on the soma and proximal dendrites of pyramidal neurons generating large sIPSCs (Freund and Katona, 2007), we next examined whether the reduction in the number of this subtype of interneurons could be related to the large sIPSC events. We performed an objective amplitude distribution analysis used in other study (Chen et al., 2018) for separating small and large events mediated GABA_A receptors, and the results obtained are consistent with the unchanged overall sIPSCs. Again, PAE did not modulate either large and small events sIPSCs suggesting that the interneuronopathy found in PPC does not affect GABA_A receptor function measured in adult offspring.

One possible explanation of the lack of significant effects may be the low average BAC (30 mg/dl) measured in the pregnant dams in these cohorts. Previously, PAE with a high average of BAC (80 mg/dl) was shown to significantly increase sIPSC frequency measured from mPFC pyramidal neurons located in layer V (Skorput et al., 2015). However, it should be noted that average intake was similar in the current study and that previous report, suggesting that changes in drinking patterns, such as front loading, may greatly affect total BAC measured across gestation (Maphis et al., 2022). In contrast, our data are in line with previous results showing a significant decrease in the average frequency and amplitude of spontaneous excitatory postsynaptic currents (sEPSCs) recorded from pyramidal neurons of mPFC layer V mPFC after PAE with low BAC (20 mg/dl), but no significantly alteration on spontaneous GABAergic transmission (Skorput and Yeh, 2016). In addition, using the current drinking model with similar drinking levels, we have recently shown that PAE mice have a long list of behavioral impairments in adulthood including altered visuospatial learning (Kenton et al., 2020), altered reinternment learning (Olguin et al., 2019), altered cognitive control and PPC beta power (Olguin et al., 2021; Olguin et al., 2023).

An overall significant reduction in PV⁺ interneurons in the PPC together with non-significant alterations of post-synaptic GABA_A receptor function in pyramidal neurons may shift the E/I balance toward excitation. We speculate that the lowered number of PV+ interneurons observed in PAE mice might increase the excitatory input to pyramidal neurons. Furthermore, this hypothesis could explain the poor response inhibition observed in PAE mice during the performance of 5C-CPT task (Olguin et al., 2021), assuming that the overall decrease of PV⁺ GABAergic interneurons increases the glutamatergic tone in PCC. However, there are several potential limitations of the current investigation that should be noted. In the current study, all mice were behavioral naïve, and it is highly possible that the PAE effect on GABAergic function could be significantly altered during the performance of behavioral tasks during which PPC function is heavily recruited. Considering previous findings that PAE impaired behavioral performance on PPC-mediated tasks and reduced neuronal activity as measured by EEG-like recordings (Olguin et al., 2023), future studies need to measure inhibitory tone following behavioral testing on PPC-mediated tasks.

In addition, the unaltered GABAergic function observed despite the significant overall reduction of PV+ interneurons in PPC could be explained by the role of astrocytes in the control of E/I balance. Considering that only ~ 20-30% of the cortical cells are interneurons (Markram et al., 2004) it is possible that interneuron-astrocyte signaling participate indirectly to regulate E/I balance through a different mechanisms (Mederos and Perea, 2019). To our knowledge there are still few studies investigating the role of astrocytes in FASD include rodent alcohol exposure paradigms. Interestingly, a briefly exposure to moderate dose of alcohol in rats from PND 5 to 7 induced a significant change in astrocyte gene expression in cortex (Fletcher and Shain 1993). Furthermore, another study using an intragastric intubation model observed a significant astrogliosis in parietal cortex (Goodlett et al., 1997). Despite the different paradigms used, these studies suggest that neonatal alcohol exposure affects the astrocytes, and it would be interesting to evaluate if in our experimental paradigm we observed long-term effects on number of astrocytes.

Overall, the current data demonstrate that alcohol exposure during the first- and second- trimester equivalents induces a significant alteration in the number of PV⁺ interneurons in PPC measured in adult offspring. They also suggest that at least at baseline conditions, changes in E/I balance may compensate for loss of inhibitory interneurons. Furthermore, these data suggest a potential mechanism by which PAE may alter cognitive control in adulthood and underline the need for future studies examining how PAE changes the response to conditions that executive control processes.