Neural correlates of learning and memory are altered by early-life stress

.


Introduction
Memory, the ability to encode, store, and retrieve information, plays a crucial role in adaptation to later life challenges (Nairne & Pandeirada, 2016).Emotionally arousing experiences are especially important for shaping learning and memory formation (Cahill & McGaugh, 1995;DiFazio et al., 2022;McGaugh & Roozendaal, 2002;Sandi, 2013).However, when these experiences are more traumatic, they can have detrimental effects on cognitive processes such as memory formation (Bolton et al., 2017;Sharpe et al., 2021).
The early postnatal time window is a critical period for neurodevelopment (Bale et al., 2010;Luby et al., 2020), and environmental factors during this sensitive period, such as early-life stress, can influence brain development and cognition later in life (Bolton et al., 2017).
Recent studies show that stress and stress hormones, such as glucocorticoids, modulate the number and activation of learning-activated cells (Brosens et al., 2023;Lesuis et al., 2021).Yet, whether and how ELS alters neuronal activation patterns is unknown.In addition, whether such ELS effects can be mitigated remains to be investigated.Interestingly, there is evidence that blocking stress signalling during a later developmental window, such as adolescence, may prevent the negative effects of ELS on cognitive function.In models involving brief maternal separation and housing mice with limited bedding and nesting material, the use of RU486, a GR antagonist, was found to prevent ELS-induced cognitive impairments in contextual memory (Arp et al., 2016;Loi et al., 2017).
Therefore, in this study we investigated the effects of ELS (induced by LBN) on contextual memory and activation of brain cells by examining expression patterns of Arc and c-Fos.We used male Arc::dVenus mice, in which the expression of a destabilized fluorescent reporter (dVenus) is linked to the Arc promoter, allowing for the detection of learning-activated cells at later time points after learning (Brosens et al., 2023;Eguchi & Yamaguchi, 2009;Lesuis et al., 2021;Rao-Ruiz et al., 2019a).We used male mice since we observed in earlier studies that early life adversity specifically affects acquisition of fear memories in male mice (Sanguino-Gómez & Krugers, 2024).Male mice were subjected to ELS using the LBN paradigm from postnatal day 2 to postnatal day 9, and received three intraperitoneal injections of RU486 during adolescence.At adult age, animals were tested in two separate cohorts.In one cohort, mice were subjected to fear conditioning (FC) and then sacrificed for analysis.In the other cohort, mice underwent both fear conditioning training and subsequent memory retrieval in the same context 24 h later.Both cohorts were assessed to determine the expression patterns of Arc::dVenus + and c-Fos + cells in the dentate gyrus and basolateral amygdala.

Animals
All mice were housed under standard conditions (20-22 • C, 40-60 % humidity) and in a light/dark cycle of 12 h (light on from 8 a.m. to 8p. m.), with continuous radio background noise provided.Standard chow and water were provided ad libitum.We used transgenic Arc::dVenus male mice (kindly provided by prof.dr.Steven Kushner, Erasmus University Rotterdam and backcrossed for more than 10 generations into a C57BL/6J background).Arc::dVenus is a mutant mice line engineered with a variant of Venus, a bright yellow fluorescent protein, fused with degradation signal sequences transforming it into a destabilized fluorophore with variable degradation rates.This destabilized (d) Venus construct is under the control of the Arc promoter (Arc::dVenus).The Arc::dVenus mice enable the identification of experience-dependent activation of neurons after 60 min and stabilized up to 24 h in the dentate gyrus (Rao-Ruiz et al., 2019a) while maximal experiencedependent expression is found between 4-6 h in the amygdala (Eguchi & Yamaguchi, 2009;Gouty-Colomer et al., 2016).
Two separate cohorts were used in this experiment.The first cohort, consisting of 11 litters with a total of 33 animals, underwent training in the FC paradigm and was sacrificed 90 min later.The second cohort, consisting of 12 litters with a total of 39 animals, underwent the same training in the FC paradigm and animales were re-exposed to the same context 24 h later, and sacrificed 90 min the task.All experiments were conducted during the light phase.All experiments were approved by the animal Welfare committee of the University of Amsterdam, under EU directive 2010/63/EU.

Breeding
In-house breeding was established to ensure a consistent perinatal environment.Six to eight-week-old female C57BL/6J mice, derived from Envigo Laboratories (Venray, Netherlands), underwent a two-week acclimatization period before breeding.Breeding pairs consisted of two females and one Arc::dVenus homozygote male mouse (Brosens et al., 2023;Eguchi & Yamaguchi, 2009;Lesuis et al., 2021;Rao-Ruiz et al., 2019a).These pairs were housed together for ten days in a conventional type II cage with standard bedding and nesting and paper enrichment to allow mating.Next, the females were kept together for another week with standard bedding and a small amount of old bedding material to keep the odour, and half of the paper cage enrichment, as well as a new 5 x 5 cm square of cotton nestlet (Tecnilab-BMI, Someren, The Netherlands).Eighteen days after the start of the breeding, prospective pregnant primiparous females were individually housed in a conventional type II cage covered with a filtertop, normal bedding, one nestlet and a small amount of old bedding material, with no paper cage enrichment and transferred to a quiet room.Dams were checked for the birth of pups each morning before 09:00 a.m.When a litter was born, the previous day was designated as PND 0.

Early life stress paradigm
At postnatal day (PND) 2, litters were weighted as well as the dam.Litters were culled to maintain six pups per litter.Litters consisting of fewer than five pups or exclusively one sex were excluded from the study.Dams and their litters were randomly allocated to either the early life stress or control group until PND9, following which all mice received uniform treatment.ELS was induced by limiting nesting and bedding material from PND2 to PND9 as described previously (Lesuis et al., 2019;Naninck et al., 2015;Rice et al., 2008;Walker et al., 2017).In this paradigm, dams were provided with reduced sawdust (33 g) and only half of the standard nestlet.Additionally, a stainless-steel grid (measuring 22 cm x 16.5 cm with a mesh size of 5 mm x 5 mm, in-house made) was positioned 1 cm above the cage floor to hinder the use of sawdust for nesting in the ELS condition.Control mice were housed with the standard quantity of sawdust (100 g) and provided a standard nestlet.Filter tops were maintained for both groups during this period and mice from both groups were undisturbed until PND9 when dam and pups were weighted.At PND9, mice were subsequently returned to standard housing conditions, which persisted until PND21, when mice were weaned.At this age, ear clips were collected for identification purposes and to determine the genotype.

Fear conditioning
Two-months old (± 2 weeks) male mice were tested in an auditory fear conditioning paradigm.To avoid (1) bias after subsequent testing order and (2) non-fear related activation of IEGs, all mice were individually housed in isolated cabinets one-week prior testing.All behavioral procedures were scheduled in the morning, occurring between 08.00 a.m. and 10.00 a.m., in order to avoid behavioral changes due to circadian-induced fluctuations in CORT levels.
The FC protocol was conducted using a computer equipped with Ethovision software (version 14.0, Noldus, The Netherlands) and mouse behavior was recorded via a camera (Basler acA1300-30gm GigE, Ahrensburg, Germany) connected to the computer.On the training day, mice were introduced into the conditioning chamber (Context A), a square chamber measuring 17 x 17 x 25 cm that was equipped with a steel grid floor connected to a shock generator.Prior to any stimuli, mice were allowed to explore this context freely for a duration of 180 s, during which white noise (1.2 kHz, 50 dB) played in the background.Exploration was followed by a Conditioned Stimulus (CS) (30 s, 2.8 kHz, 82 dB) paired to a foot shock (Unconditioned Stimulus-US) during the two last seconds of the tone (2 s, 0.4 mA).This tone-shock pairing was presented three times, each separated by an intertrial interval of 60 s, and after the last one, animals were returned to their homecage.After each trial, the conditioning chamber was cleaned with 96 % ethanol to ensure no residual odours from previous trials influenced subsequent animals.The first cohort was sacrificed through decapitation 90 minutes after training.For the second cohort, twenty-four hours later, mice were reintroduced into Context A, where white noise at 50 dB played in the background for 180 s.Following contextual retrieval, the mice were returned to their standard housing conditions and 90 min later they were sacrificed through rapid decapitation.The timepoint of 90 min was chosen because it corresponds to the peak expression of c-Fos (Denny et al., 2014).Trunk blood was collected and brains were dissected.The left hemispheres were post-fixed in a solution of 4 % paraformaldehyde in phosphate buffer (PB 0.1 M, pH 7.4) at 4 • C and the right hemispheres were snap-frozen for further analysis.

Behavioral analysis
Throughout the experiments, freezing behavior defined as 'the absence of movement except the ones required for breathing' (Zhou et al., 2009) was scored using an automatic procedure, consisting of a deep-learning pose estimation algorithm, Deeplabcut (Mathis et al., 2018), in combination with a machine-learning tool for supervised behavior classification, SimBA (Goodwin et al., 2024).Behavior during training was scored during exploration (first 180 s) the three tone-shock pairings (30 s per period) and the intertrial intervals (60 s).Freezing during context retrieval was scored twenty-four hours after training for 180 s.Additionally, locomotor activity was determined by calculating the cumulative distance covered by the centroid point of the animals, which was tracked using Deeplabcut (Mathis et al., 2018) and transformed from pixels to mm.

Plasma corticosterone measurements
Blood samples were collected in ice cold, EDTA-coated tubes (Sarstedt, Etten-Leur, The Netherlands), placed on ice and centrifuged at 14.000 rpm for 25 min.This process isolated plasma, which was stored at − 20 • C until further analysis.Plasma corticosterone levels were measured using a commercially available test kit (Tecan, Hamburg, Germany).

Sectioning
Brains were dissected, stored, cryoprotected, and coronally sectioned for subsequent analysis.Specific details of this procedure were as follows.Twenty-four hours after dissection, the brains were stored in Phosphate Buffered Saline (PBS, 0.05 M Phosphate Buffer (PB) with 0.9 % NaCl) containing 0.01 % sodium azide at 4 • C until further use.One day before sectioning, cryoprotection took place, involving immersion in a 15 % sucrose solution in 0.1 M PB until sinking (typically taking approximately 3 h) followed by immersion in 30 % sucrose in 0.1 M PB overnight.The left hemisphere was sectioned coronally at 40 μm thickness in 6 parallel series at an angle of 3 • with a Leica Jung HN microtome.Sections were preserved in an anti-freeze solution (30 % ethylene glycol, 20 % glycerol, 50 % 0.05 M PB) at − 20 • C until further use.

Immunostaining
Sections that included the hippocampus and/or amygdala (bregma points − 0.58 to − 3.80) were selected for the immunostaining.All immunostaining procedures were carried out exclusively on the left hemisphere (Klur et al., 2009).Sections were first washed three times with PBS for 5 min.They were then incubated for 30 min with 500 μL of Fab-fragments (1:200; Goat Anti Mouse IgG Fc-BIOT, SothernBiotech (1033-08)) in PBS on a rotating table at room temperature.After incubation, sections were washed 3 times for 5 min with PBS.Next, the slices were incubated in 5 % BSA in PBST (PBS with 0.3 % Triton) for 2 h on a rotating table at room temperature.Without washing, the sections were incubated with rat anti-c-Fos (1:500; Synaptic Systems (226 017)), chicken anti-GFP (1:750; Abcam (ab13970)) and, to outline the BLA, mouse anti-GAD67 clone 1G10.2 (1:1000, Merk Millipore (MAB5406)) in 5 % BSA in PBST.Sections were incubated for one hour on a rotating table at room temperature and overnight on a rotating table at 4 • C. The next day, the sections were washed 3 times for 10 min with PBST.After washing, the sections were incubated with Alexa Fluor 568 goat anti-rat (1:200; Invitrogen, life technology cooperation Eugene (# A-11077)), Alexa Fluor 488 goat anti-chicken (1:500; Invitrogen, life technology cooperation Eugene (#A-11039)) and Alexa Fluor 647 donkey antimouse (1:500; Invitrogen, life technology cooperation Eugene (#A-31571)) and DAPI (1:1000, Vector Laboratories Inc, H-1200) in 5 % BSA in PBST for 2 h on a rotating table at room temperature.Finally, sections were washed 3 times for 5 min with PBS.The sections were then mounted using embedding medium (70 % glycerol in Tris HCl pH = with 0.5 % n-propylgallate as anti-fading agent).

Imaging and quantification
Images were obtained using a Nikon DS-Ri2 fluorescence microscope (software NIS element, version BR 4.60.0064-bit).Images were acquired at 10x magnification for the BLA and 20x for the DG with the corresponding channels (GFP (Arc::dVenus), Tx Red (c-Fos), Cy5 (GAD65), and DAPI) and later assigned to corresponding bregma points.Microscope settings were kept consistent for all scanned slices.Analysis was done blind for the experimental conditions and in random order to prevent any bias.Regions of interest (ROIs) were selected manually based on the mouse brain atlas (Paxinos and Franklin 2019) using ImageJ analysis software (version 1.53 k).For the hippocampus, the dentate gyrus was divided into dorsal (bregma points − 1.22 to − 2.18) and ventral hippocampus (bregma points − 2.30-to − 3.80).The basolateral amygdala (bregma points − 0.82 to − 1.82 mm) was outlined using GAD67 staining as counterstaining (Supplementary Figure 1).Images were converted to 8-bit gray scale to visualize positive cells.Since the background did not always have the same intensity in the whole image, a ratio of grey values for the cell intensity/background of 1.05 was used.Cells were considered positive for Arc::dVenus or c-Fos when the grey value of the cell was above 1.05 x the grey value of the background.A representative image demonstrating colocalization is shown in Supplementary Figure 2. Counting was done manually, and the absolute values were converted to the number of cells per mm 2 .For visualization and analysis purposes, cell counts were averaged across all slices for each area per animal.

Statistical analysis
Before the start of the experiments, animals were allocated randomly to their respective conditions and treatments, and data analysis was carried out by an experimenter who remained blinded to the experimental conditions throughout the analysis process.To determine the appropriate sample size, a priori power analysis was conducted using GPower 3.1.For statistical analysis and data visualization, we utilized R version 4.2.1 (2022-06-23 ucrt) and RStudio version (2022.7.2.576).Details of the packages employed and their corresponding citations can be found in Supplementary Table 1.Outliers were identified using the interquartile range (IQR) method: data points that fell either above Q3 + 3*IQR or below Q1 − 3*IQR were considered as outliers.Any outliers that were detected and found to be attributable to methodological or biological irregularities were subsequently removed from the dataset.If an outlier was removed for one variable, it was excluded from the entire analysis.Conversely, owing to technical constraints, specific animals lacked complete data for all brain areas, leading to slight variations in the final animal numbers for each analysis.
The normality of the data was assesed using the Shapiro-Wilk test and homogeneity of variance through Levene's test.Sphericity was assessed using Mauchly test and whenever data didn't meet these criteria, Greenhouse-Geisser corrected results were used.Given the nature of the experiment, the ELS assignment was intrinsically linked to specific litters.Therefore, before conducting any analyses, we examined nested effects by conducting an analysis of variance (ANOVA) that compared the use of litter as a random factor.When this test was significant, the data was analysed including the litter as a random factor in a linear mixed model (Aarts et al., 2014).
For single-factor parameters, unpaired t-tests were employed.Parameters with two factors underwent a two-way ANOVA with "ELS" and "RU486 treatment" as the between factors.Parameters with three factors underwent a two-way ANOVA with "ELS" and "RU486 treatment" and "task" as the between factors.For the remaining analyses, a 3 x 2 x repeated measures factorial ANOVA was carried out to evaluate the interaction between the between-subject factors (ELS and RU486 treatment) and the within-subject factor (time).To accomplish this, a mixed model ANOVA with two between-subject factors (ELS and RU486 treatment) and one within-subject factor (time) was used.Mixed model was used even though the full data set didn't meet the assumption of normality, since the mixed model is quite robust to violations of this assumption.In cases where interaction effects were observed, post-hoc analyses were conducted.Partial eta-squared (ηp2) was used to assess effect size.A significance level of p ≤ 0.05 was considered statistically significant.When necessary, adjustments were made to the p-value using Bonferroni correction to control for multiple comparisons.

Body weight gain
Exposure to limited nesting and bedding material (LBN) conditions between PND 2 and PND 9 reduced body weight gain (t(14.72)= 3.402, p = 0.004, Table 1).Body weight in of ELS mice returned to control levels after returning the mice to their standard housing conditions at PND 9 (body gain from PND 9 till adolescence − average PND 28-30: t (13.22) = 0.438, p = 0.669).For absolute body weights at PND 2, PND and adolescent age, see Supplementary Table 2).
In a separate cohort, male mice were trained and subsequently assessed for contextual memory twenty-four hours later by measuring freezing behavior as a percentage of time spent freezing during 3 min of re-exposure to the training context.ELS reduced freezing levels (F(1,35) = 4.513, p = 0.041, Fig. 1E), with RU486 treatment during adolescence not modifying the impact of ELS on freezing behavior.
To account for potential effects of general locomotor activity on freezing levels, we measured the distance travelled during the exploration period preceding tone/footshock application in cohort 1 and the distance travelled during the whole contextual retrieval task in cohort (Fig. 1C,F).Notably, locomotor activity was unaffected by both ELS and RU486 treatment.
Finally, plasma corticosterone levels collected 90 min after the task (either training or contextual retrieval) did not exhibit any significant differences between experimental groups (Fig. 1D,G).

IEGs after training in the DG
To dissect the effects of ELS on cellular activity during memory acquisition, we examined activation after training in the DG and BLA, two brain regions strongly involved in auditory and contextual fear memory (Phillips & LeDoux, 1992).We therefore used the Arc::dVenus transgenic mouse line (Eguchi & Yamaguchi, 2009;Lesuis et al., 2021;Rao-Ruiz, et al., 2019) characterized by destabilized Venus fluorescent protein expression under the control of the activity-regulated cytoskeleton-associated protein Arc promoter, in combination with an immunostaining against c-Fos.Arc (and consequently, Arc::dVenus) expression is primarily found in excitatory neurons involved in synaptic plasticity and memory processing whereas c-Fos is considered a general stimuli-activated marker expressed in various brain cell types, including excitatory and inhibitory neurons and glial cells (Condorelli et al., 1989;  Cruz-Mendoza et al., 2022;Lukkes et al., 2012;Staiger et al., 2002).
As ELS constitutes a developmental manipulation, and despite normalizing the IEG count, we checked the DG surface area which was not altered by ELS (Supplementary Figure 3).Ninety minutes after training, we observed a reduction in the number of Arc::dVenus + cells in the DG of ELS-exposed mice both in the dorsal (F(1,29) = 10.879,p = 0.003, Fig. 2A) and ventral regions (F(1,29) = 12.105, p = 0.002, Fig. 2F) of the DG.No effects of RU486 were found.No effect of ELS nor RU486 was observed for the number of c-Fos + cells (Fig. 2B,G) or the number of double-labeled Arc::dVenus + and c-Fos + cells (Fig. 2C,H).However, an increase in percentage of double-labeled cells within the total Arc::dVenus + population was observed in the ELS group both in dorsal (F(1, 29) = 7.597, p = 0.010, Fig. 2D) and ventral (F(1, 29) = 5.519, p = 0.003, Fig. 2I) DG.The percentage of double-labeled cells within the total c-Fos + population was decreased by ELS exclusively in the dorsal DG (F(1, 29) = 7.654, p = 0.0010, Fig. 2E).Representative pictures are shown in Fig. 2K.

IEGs after retrieval in the DG
Ninety minutes after retrieval, ELS decreased the number of c-Fos + cells in the ventral part of the DG (F(1, 33) = 8.296, p = 0.007, Fig. 3E), and reduced the number of Arc::dVenus + cells that also expressed c-Fos (F(1, 33) = 5.237, p = 0.029, Fig. 3F).No effects were found in the dorsal DG (Fig. 3B,C).There was no effect in the number of Arc:: dVenus + cells (Fig. 3A, D) or in the percentages of colocalized cells with the total Arc::dVenus + or c-Fos + populations (Fig. 3D, E, I, J).Additionally, there were no effects of RU486 observed in either of the activation markers.Representative pictures are shown in Fig. 3K.

Changes in IEGs between training and retrieval in the DG
We next assessed whether the expression of IEGs in the DG changed differently among groups between training and retrieval.It is worth to mention that two separate cohorts were analysed with this purpose and therefore we lack real temporal dynamics.Also, the two cohorts were stained separately.We therefore conducted a pilot study (Supplementary Table 3), and stained control samples from both timepoints using the same method.The results were consistent with those when stained separately suggesting that the results of the stainings are comparable.

IEGs after training in the BLA
Similar to the DG, a decrease in the number of Arc::dVenus + cells was observed in the BLA of the ELS groups (F(1,28) = 10.181,p = 0.003, Fig. 5A) ninety minutes after acquisition.Furthermore, a reduction in the number of colocalized c-Fos and Arc::dVenus cells was noted (F (1,28) = 7.574, p = 0.010, Fig. 5C).Further analysis found a decrease in the percentage of Arc::dVenus + cells among the total c-Fos + population (F(1, 28) = 12.206, p = 0.002, Fig. 5E).No effects of RU486 were observed in any of these measures.Representative pictures are shown in Fig. 5G.No changes in BLA surface area after ELS were observed for any timepoint (Supplementary Figure 3).

IEGs after retrieval in the BLA
Comparable to the DG, ELS decreased the number of c-Fos + cells in the BLA ninety minutes after contextual retrieval (F(1,32) = 7.699, p = 0.003, Fig. 6B).No effects of RU486 were observed in any of these measures.Representative pictures are shown in Fig. 6G.

Discussion
The present study investigated the impact of ELS induced by housing dams and offspring with limited bedding and nesting material between PND 2-9 on memory formation in male mice and whether the glucocorticoid receptor antagonist, RU486, could mitigate these effects.Furthermore, we studied the effects of ELS and RU486 treatment on brain activation by the expression of the IEGs c-Fos and the fluorescencebased Arc reporter Arc::dVenus after both training and contextual retrieval.We observed that ELS reduced both the acquisition and contextual memory retrieval at 24 h after training.These effects were not mitigated by treatment with the GR antagonist RU486 at adolescence.Immediately after training, ELS reduced the number of Arc:: dVenus + cells, but not the number of c-Fos + cells, in both DG and BLA.
After retrieval, ELS decreased the number of c-Fos + cells in the ventral DG and BLA and decreased of the number of co-localized c-Fos and Arc:: dVenus cells in the ventral DG.In addition, ELS altered IEGs expression over time and increased coactivation patterns upon training.These results suggest ELS disrupts memory acquisition and retrieval and alters neural activation both during acquisition and retrieval.Blocking glucocorticoid receptors during adolescence did not mitigate these effects.

Early-life stress impairs both fear acquisition and contextual retrieval
Various studies have demonstrated effects of ELS, induced by LBN, on spatial, contextual and emotional memory (Brunson et al., 2005;Chocyk et al., 2014;Graham, 2018;Kosten et al., 2006;Lesuis et al., 2019;Naninck et al., 2015;Pillai et al., 2018;Rice et al., 2008;Wang et al., 2012).Here we explored the effects of ELS not only on memory but also on fear memory acquisition.In our study, we observed that ELS not only impaired contextual memory but also significantly hampered memory acquisition.This finding aligns with prior work, including Stevenson et al. (2009), Bordes et al., (2024) and our own previous studies (Sanguino-Gomez & Krugers, 2024).Importantly, these learning and memory deficits do not appear to derive from ELS effects on locomotor activity as, in our study, the latter was unaffected by ELS.These observations suggest that the observed behavioral changes are predominantly linked to cognitive deficits rather than physical reactions to the aversive stimulus.Targeting GRs using RU486 at adolescence age was during adulthood at present not effective to prevent ELS effects on acquisition or retrieval.This contrasts with earlier research reporting that blocking GRs during adolescence could effectively mitigate ELS-induced learning and memory deficits (Arp et al., 2016;Loi et al., 2017;Sanguino-Gómez & Krugers, 2024).One possible explanation for this discrepancy is that the preventive effect of targeting GRs may be contingent on repeated exposure to memory test sessions, alike in prior studies (Arp et al., 2016;Loi et al., 2017) but may also be due other interstudy differences and/or lack of statistical power in our present study.
Finally, there was no impact of ELS on plasma CORT levels, which are involved in memory regulation (Brosens et al., 2023;dos Santos Corrêa et al., 2021;Lesuis et al., 2021).In particular, ELS did not change CORT levels following a second stressor (fear conditioning) which is in line with earlier studies (for an extensive review, (van Bodegom et al., 2017).We also found that plasma corticosterone levels did not correlate with freezing levels, suggesting that circulating CORT levels in our study are not a major determinant for memory expression.Whether the absence of an ELS effect on corticosterone levels is due to the chosen time point (90 min after training or retrievalwhen the expected peak in CORT levels has passed − Spencer & Deak, 2017) remains to be investigated.

Early-life stress decreases brain activation: Arc and c-Fos are differently regulated after training and retrieval
Various lines of evidence have provided evidence that expression of IEGs, including the proto-oncogene c-Fos and the activity-regulated cytoskeleton-associated protein Arc, reflect experience-dependent and learning induced activation of cells in the brain (Gouty-Colomer et al., 2016;Guzowski et al., 2001;Mahringer et al., 2019).Gain-on and gainoff studies have shown that that these activated cells play a pivotal role in (fear) memory formation (Garner et al., 2012;Han et al., 2009;Liu et al., 2012;Ramirez et al., 2013).
Here we measured the number of c-Fos and Arc::dVenus expressing neurons as an index of brain activation after training and contextual memory retrieval.Our current IEG studies using contextual recall provided only a partial reexposure to the entire training experience, that included also cued tone as a predictor.Simultaneous exposure to both cue recall and contextual recall might limit our ability to directly connect cue memory with IEG expression.We therefore focused in this study on contextual memory.
The absolute number of c-Fos + and Arc::dVenus + cells in DG and BLA after training and retrieval in control animals was in line with findings from other studies (Brosens et al., 2023;Matsuo, 2009;Silva et al., 2019).However, the percentage of colocalization between Arc::dVenus (caption on next column) Fig. 7. ELS differentially modifies the changes in the cell percentages between training and retrieval.A. Barplots showing the number of Arc::dVenus + cells per mm 2 (mean ± SEM) in the BLA 90 min after FC versus 90 min after reexposure to the context 24 h later.Retrieval increased Arc::dVenus expression.B. Barplots showing the number of c-Fos + cells per mm 2 (mean ± SEM) in the BLA 90 min after FC versus 90 min after reexposure to the context 24 h later.Retrieval decreased the number of c-Fos + cells.C. Barplots showing the number of double labeled cells per mm 2 (mean ± SEM) in the BLA 90 min after FC versus 90 min after reexposure to the context 24 h later.Expression was decreased upon retrieval.D. Barplots showing the percentage of c-Fos + cells among the total Arc::dVenus + population per mm 2 (mean % ± SEM) in the BLA 90 min after FC versus 90 min after reexposure to the context 24 h later.Retrieval reduced this percentage.E. Barplots showing the percentage of Arc:: dVenus + cells among the total c-Fos + population per mm 2 (mean % ± SEM) in the BLA 90 min after FC versus 90 min after reexposure to the context 24 h later.This percentage was increased upon retrieval specifically in the ELS and c-Fos expression was lower as reported by others (Lesuis et al., 2021), which may be due to detection methodologies such as the threshold that was chosen.
Following training, ELS decreased the number of Arc::dVenus + cells in both dorsal and ventral DG as well as the BLA.It is worth noting that Arc::dVenus allows for the visualization of active cells up to 24 h prior in the DG and approximately up to 6 h in the amygdala Eguchi & Yamaguchi, 2009;Gouty-Colomer et al., 2016;Rao-Ruiz et al., 2019a).Therefore, one might argue that the effects of ELS on the Arc::dVenus + population could be attributed to alterations in baseline activity.However, we found no effect of ELS on the number of c-Fos + and Arc:: dVenus + cells at baseline (without training, unpublished observations).
Since Arc is predominantly present in excitatory cells (Moga et al., 2004) while c-Fos is expressed also in other cell types (Condorelli et al., 1989;Cruz-Mendoza et al., 2022;Lukkes et al., 2012;Staiger et al., 2002), the absence of a decrease in the number of c-Fos + cells after training may suggest the activation of other cell-types, such as glial cells or inhibitory neurons in ELS animals.To further investigate how ELS affects cell-type-specific activation after fear conditioning, double labeling of c-Fos with microglia, astrocytes and inhibitory markers will be important, which was beyond the scope of the current paper.
Following retrieval, ELS decreased the number of c-Fos + cells specifically in the ventral DG as well as in the BLA while no changes in the number of Arc::dVenus + cells were found.The reduction in number of c-Fos + cells may suggest a general cell-type reduced activation after retrieval in ELS mice, rather than a specific excitatory, possibly plasticity related change.
In the DG, where the Arc::dVenus signal persists for up to 24 h, this reporter can serve as a valuable tool for examining the cells activated during recall (c-Fos) that were previously active during training (Arc:: dVenus).Notably, the number of these double-labeled cells was reduced by ELS in the ventral DG.This reduction suggests that the ensemble or engram formed to represent the fear conditioning (FC) memory may be compromised by ELS, potentially contributing to the mechanism behind the observed memory deficits.While sparsity is essential for effective engram formation and memory encoding (Ramsaran et al., 2023;Rao-Ruiz et al., 2019b), excessive sparsity due to ELS may be detrimental for engram allocation, as it may impede the establishment of synaptic connections, ultimately leading to memory impairments.However, gain and loss of function studies on the engram would be essential to establish causality (Garner et al., 2012;Han et al., 2009;Liu et al., 2012;Ramirez et al., 2013).A similar trend towards a decrease in double-labeling was found in the BLA.Since Arc::dVenus is not expressed for 24 h in the amygdala (Rao-Ruiz et al., 2019a), it is not possible to link its expression to training activated cells.Further research will be necessary, and the use of tools enabling more enduring labeling of activated cells, such as the TRAP2 line (DeNardo et al., 2019), may offer further insights of how cell activation in the BLA contributes to the memory changes induced by ELS.
The observed changes in cellular activation after retrieval predominantly occurred in the more ventral region of the DG, which is more connected to the amygdala and hypothalamus and which has been implicated in anxiety, fear, and stress responses (Fournier & Duman, 2013;McHugh et al., 2004).This region is also highly sensitive to ELS (Fulton et al., 2021;Grigoryan & Segal, 2016;Murthy et al., 2019).Moreover, changes in ventral DG structure have been linked to an increased susceptibility to trauma (Dirven et al., 2022).

ELS affects over time dynamics of IEGs expression
Next, we compared the number of endogenously expressing Arc:: dVenus + cells and c-Fos + cells between training and retrieval.Staining for Arc::dVenus en c-Fos was done on different cohorts which warrants careful interpretation of this comparison.Also, stainings per experiment were conducted at different time points, although a pilot study where sections collected after training and retrieval were stained confirmed the experimental data.Over time, between training and retrieval, we observed an increase in the number of endogenously expressing Arc:: dVenus + cells in the dorsal DG in ELS animals.Since control animals express the same number of Arc::dVenus + cells after training and retrieval, there might be (compensatory) normalization of the ELS effects on Arc::dVenus cells, suggesting recruitment of Arc::dVenus + cells after training in DG.This may have occurred during consolidation and/ or after retrieval due to the dynamics of Arc::dVenus expression.The latter is more likely to have occurred in the BLA, given that the Arc:: dVenus signal is only present during a limited period after experiencedependent activation in this area.While comparing, this increase in Arc::dVenus cells over time was more prominent in the basolateral amygdala than in DG (effect size ηp 2 = 0.508 in BLA versus ηp 2 = 0.132 in the DG), suggesting that the that BLA of ELS animals engages a stronger excitatory population during contextual retrieval which may be relevant with its role in fear and anxiety (Davis, 1992;Ressler, 2010).The finding that ELS increases the number of Arc::dVenus + cells after retrieval in the DG raises the possibility that ELS memory deficits observed during retrieval are a consequence of (delayed) acquisition or encoding problems, as previously hypothesized (Sanguino-Gómez & Krugers, 2024).
Since Arc::dVenus in the DG can stay up to 24 h after experience, pinpointing the exact timing of this activation for normalization remains unknown.It could be that during recall, the excitatory population is overactivated, or that this increase in the number of Arc::dVenus + cells occurs during consolidation and may take place later than 90 min after training to visualize it.In the latter case ELS may have delayed the ability for learning-induced plasticity.
Regarding c-Fos expression, we observed a lower general activation upon recall when compared to the initial training phase in control and ELS animals.This observation suggests that the initial tone-footshock pairings trigger a larger activation compared to mere contextual reexposure.ELS did not significantly alter the pattern over the different memory phases.Likewise, the number of colocalized cells remained approximately the same over time in the DG but in the BLA these levels decrease during retrieval.
In our study, the number of Arc::dVenus + cells that was also c-Fos + was decreased in the BLA of ELS animals.However, in this area, ELS animals have an increased percentage of Arc::dVenus + cells within the total c-Fos population when comparing training and retrieval phases (3-4 % versus 15-18 % respectively).This shift may possibly reflect altered activation of Arc::dVenus + cells possibly at the expense of altered activation in c-Fos in other cell types.Whether these changes reflect changes in excitatory/inhibitory balance remains to be studied but could be important for maintaining the sparsity of the engram (Chen et al., 2022;Guo et al., 2018;Pignatelli et al., 2019;Ramsaran et al., 2023), a key property of these memory ensembles (Josselyn et al., 2015;Tonegawa et al., 2015).

Conclusion
Here we report that ELS, induced by housing dam and offspring with limited bedding and nesting material from PND2-9, impairs acquisition and retrieval of fear memories.Administration of the glucocorticoid receptor antagonist RU486 during adolescence did not mitigate the memory deficits induced by ELS.Analysis of IEG expression after training and upon retrieval revealed distinct alterations in activation within the DG and BLA.After training, ELS reduced the number of Arc:: dVenus + cells, possibly indicating reduction in synaptic plasticity and the encoding of memories, particularly within the population of excitatory cells.However, the number of c-Fos + cells remained unaffected, which may suggest that overall neural activation during this phase was not compromised.In contrast, after memory retrieval, ELS decreased the number of c-Fos + cells in the ventral DG and BLA, indicating reduced general neural activation during recall.The Arc::dVenus population remained unaffected upon recall, possibly reflecting a compensatory mechanism within the population of excitatory cells and a potential reduction in the activity of other cell types besides excitatory neurons.Colocalization in the ventral DG upon retrieval was decreased on ELS, suggesting an impairment in engram cell allocation.Changes over time were also altered due to ELS.These results shed light on the potential mechanisms underlying the memory deficits induced by ELS through changes on its neural corelates.While these findings provide valuable insights, further research using novel experimental tools is imperative to fully comprehend the intricate processes underlying these changes including causal studies, cell-type-specific activation and compensatory mechanisms.Ultimately, these findings may offer potential avenues of intervention strategies for preventing ELS-induced cognitive deficits.

Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Fig. 1 .
Fig. 1.ELS decreased acquisition and retrieval of fearful memories without affecting locomotor activity or plasma CORT levels.A. Schematic overview of the auditory fear conditioning experiment.Male mice were subjected to either standard conditions or ELS from PND 2-9.During adolescence (PND 28-30), mice received injections with vehicle or the glucocorticoid receptor (GR) antagonist RU486 (10 mg/kg).At PND 60, mice underwent training in a mild auditory fear conditioning task.24 h later, their contextual memory was assessed by reexposing them to the same context.Created with BioRender.com.B. Lineplots showing freezing (mean % ± SEM) across the course of conditioning.Exp = exploration, S = shock, Int = intertrial period.ELS reduced freezing acquisition over conditioning.C. Barplots showing distance in mm (mean % ± SEM) travelled during the exploration time on the training day.Distance travelled was comparable between experimental groups regardless of stress and treatment condition.D. Barplots showing plasma corticosterone circulating levels (in ng/ml, mean % ± SEM) 90 min after fear conditioning.No changes were found due to ELS or treatment.E. Barplots showing freezing (mean % ± SEM) when reexposure to training context 24 h after fear conditioning.ELS animals froze less than control mice.F. Barplots showing distance in mm (mean % ± SEM) travelled during reexposure to the conditioning context 24 h later.Distance was comparable regardless of stress and treatment condition.G. Barplots showing plasma corticosterone circulating levels (in ng/ml, mean % ± SEM) 90 min after reexposure to the conditioning context.No changes were found due to ELS or treatment.Training: N Control-Vehicle = 8, N Control-RU486 = 8, N ELS-Vehicle = 9, N ELS-RU486 = 8.Retrieval: N Control-Vehicle = 11, N Control-RU486 = 13, N ELS-Vehicle = 7, N ELS-RU486 = 8.*: ELS effect.* p ≤ 0.05.

FFig. 2 .
Fig. 2. ELS decreased the number of Arc::dVenus + cells in dorsal and ventral DG after training.A. Barplots showing the number of Arc::dVenus + cells per mm 2 (mean ± SEM) in the dorsal DG 90 min after FC.ELS reduced the Arc::dVenus + population.B. Barplots showing the number of c-Fos + cells per mm 2 (mean ± SEM) in the dorsal DG 90 min after FC.No changes were found due to ELS or RU486 treatment.C. Barplots showing the number of double labeled Arc::dVenus + and c-Fos + cells per mm 2 (mean ± SEM) in the dorsal DG 90 min after FC.No changes were found due to ELS or RU486 treatment.D. Barplots showing the percentage of c-Fos + cells among the total Arc::dVenus + population per mm 2 (mean % ± SEM) in the dorsal DG 90 min after FC.ELS animals had an increased percentage in comparison with controls.E. Barplots showing the percentage of Arc::dVenus + cells among the total c-Fos + population per mm 2 (mean % ± SEM) in the dorsal DG 90 min after FC.ELS animals had a decreased percentage with respect to controls.F. Barplots showing the number of Arc::dVenus + cells per mm 2 (mean ± SEM) in the ventral DG 90 min after FC.ELS reduced the Arc::dVenus + population.G. Barplots showing the number of c-Fos + cells per mm 2 (mean ± SEM) in the ventral DG 90 min after FC.No changes were found due to ELS or RU486 treatment.H. Barplots showing the number of double labeled Arc::dVenus + and c-Fos + cells per mm 2 (mean ± SEM) in the ventral DG 90 min after FC.No changes were found due to ELS or RU486 treatment.I. Barplots showing the percentage of c-Fos + cells among the total Arc::dVenus + population per mm 2 (mean % ± SEM) in the ventral DG 90 min after FC.ELS animals had an increased percentage in comparison with controls.J. Barplots showing the percentage of Arc::dVenus + cells among the total c-Fos + population per mm 2 (mean % ± SEM) in the ventral DG 90 min after FC.No changes were found between experimental conditions.K. Representative pictures depleting DAPI, endogenous Arc::dVenus signal enhanced by GFP and c-Fos, as well as colocalization between Arc::dVenus and c-Fos (Merge).Scale bar = 100 μm.N Control-Vehicle = 8, N Control-RU486 = 8, N ELS-Vehicle = 9, N ELS-RU486 = 8. *: ELS effect.* p ≤ 0.05, ** p ≤0.01.

Fig. 3 .
Fig.3.ELS decreased the number of c-Fos + and colocalized cells in ventral DG after contextual retrieval.A. Barplots showing the number of Arc::dVenus + cells per mm 2 (mean ± SEM) in the dorsal DG 90 min after reexposure to the context 24 h later.There were no changes between experimental groups.B. Barplots showing the number of c-Fos + cells per mm 2 (mean ± SEM) in the dorsal DG 90 min after reexposure to the context 24 h later.No changes were found due to ELS neither to RU486 treatment.C. Barplots showing the number of double labeled Arc::dVenus + and c-Fos + cells per mm 2 (mean ± SEM) in the dorsal DG 90 min after reexposure to the context 24 h later.No changes were found due to ELS or RU486 treatment.D. Barplots showing the percentage of c-Fos + cells among the total Arc::dVenus + population per mm 2 (mean % ± SEM) in the dorsal DG 90 min after reexposure to the context 24 h later.There were no changes between experimental groups.E. Barplots showing the percentage of Arc::dVenus + cells among the total c-Fos + population per mm 2 (mean % ± SEM) in the dorsal DG 90 min after reexposure to the context 24 h later.No changes were found due to ELS or RU486 treatment.F. Barplots showing the number of Arc::dVenus + cells per mm 2 (mean ± SEM) in the ventral DG 90 min after reexposure to the context 24 h later.There were no changes between experimental conditions.G. Barplots showing the number of c-Fos + cells per mm 2 (mean ± SEM) in the ventral DG 90 min after reexposure to the context 24 h later.ELS reduced the c-Fos + population.H. Barplots showing the number of double labeled Arc::dVenus + and c-Fos + cells per mm 2 (mean ± SEM) in the ventral DG 90 min after reexposure to the context 24 h later.ELS decreases the colocalization compared to controls I. Barplots showing the percentage of c-Fos + cells among the total Arc::dVenus + population per mm 2 (mean % ± SEM) in the ventral DG 90 min after reexposure to the context 24 h later.There were no changes between experimental groups.J. Barplots showing the percentage of Arc:: dVenus + cells among the total c-Fos + population per mm 2 (mean % ± SEM) in the ventral DG 90 min after reexposure to the context 24 h later.No changes were found due to ELS or RU486 treatment.K. Representative pictures depleting DAPI, endogenous Arc::dVenus signal enhanced by GFP and c-Fos, as well as colocalization between Arc::dVenus and c-Fos (Merge).Scale bar = 100 μm.N Control-Vehicle = 10, N Control-RU486 = 13, N ELS-Vehicle = 7, N ELS-RU486 = 8.*: ELS effect.* p ≤ 0.05, ** p ≤ 0.01.

Fig. 4 .
Fig. 4. ELS differentially modified the changes in the cell percentages between training and retrieval.A. Barplots showing the number of Arc::dVenus + cells per mm (mean ± SEM) in the dorsal DG 90 min after FC versus 90 min after reexposure to the context 24 h later.Retrieval increased Arc::dVenus expression, but this is almost significantly exclusive to ELS animals.Overall, ELS decreased Arc::dVenus expression.B. Barplots showing the number of c-Fos + cells per mm 2 (mean ± SEM) in the dorsal DG 90 min after FC versus 90 min after reexposure to the context 24 h later.Retrieval decreased the number of c-Fos + cells.Overall, ELS decreases c-Fos expression.C. Barplots showing the number of double labeled cells per mm 2 (mean ± SEM) in the dorsal DG 90 min after FC versus 90 min after reexposure to the context 24 h later.No changes were found in training compared to retrieval.D. Barplots showing the percentage of c-Fos + cells among the total Arc::dVenus + population per mm 2 (mean % ± SEM) in the dorsal DG 90 min after FC versus 90 min after reexposure to the context 24 h later.Retrieval decreased this percentage and ELS exclusively modifies training percentage.E. Barplots showing the percentage of Arc::dVenus + cells among the total c-Fos + population per mm 2 (mean % ± SEM) in the dorsal DG 90 min after FC versus 90 min after reexposure to the context 24 h later.The percentage was increased upon retrieval.F. Barplots showing the number of Arc::dVenus + cells per mm 2 (mean ± SEM) in the ventral DG 90 min after FC versus 90 min after reexposure to the context 24 h later.Retrieval increased Arc::dVenus expression.G. Barplots showing the number of c-Fos + cells per mm 2 (mean ± SEM) in the ventral DG 90 min after FC versus 90 min after reexposure to the context 24 h later.Retrieval decreased the number of c-Fos + cells.H. Barplots showing the number of double labeled cells per mm 2 (mean ± SEM) in the ventral DG 90 min after FC versus 90 min after reexposure to the context 24 h later.There were no differences in number between training and retrieval.I. Barplots showing the percentage of c-Fos + cells among the total Arc::dVenus + population per mm 2 (mean % ± SEM) in the ventral DG 90 min after FC versus 90 min after reexposure to the context 24 h later.Retrieval decreased this percentage and ELS exclusively increases training percentage.J. Barplots showing the percentage of Arc::dVenus + cells among the total c-Fos + population per mm 2 (mean % ± SEM) in the ventral DG 90 min after FC versus 90 min after reexposure to the context 24 h later.The percentage was increased upon retrieval.Training: N Control-Vehicle = 8, N Control-RU486 = 8, N ELS-Vehicle = 9, N ELS-RU486 = 8.Retrieval: N Control-Vehicle = 10, N Control-RU486 = 13, N ELS-Vehicle = 7, N ELS-RU486 = 8.*: ELS effect.*:ELS effect, &: time effect.Effect p ≤ 0.05.

Fig. 5 .
Fig. 5. ELS decreasesd the number of Arc::dVenus + and double labeled c-Fos and Arc::dVenus cells in the BLA after training.A. Barplots showing the number of Arc:: dVenus + cells per mm 2 (mean ± SEM) in the BLA 90 min after FC.ELS reduced the Arc::dVenus + population.B. Barplots showing the number of c-Fos + cells per mm (mean ± SEM) in the BLA 90 min after FC.No changes were found due to ELS or RU486 treatment.C. Barplots showing the number of double labeled Arc::dVenus + and c-Fos + cells per mm 2 (mean ± SEM) in the BLA minutes after FC.ELS reduced the number of colocalized cells respect to controls.D. Barplots showing the percentage of c-Fos + cells among the total Arc::dVenus + population per mm 2 (mean % ± SEM) in the BLA 90 min after FC.There were no changes due to ELS or RU486 treatment.E. Barplots showing the percentage of Arc::dVenus + cells among the total c-Fos + population per mm 2 (mean % ± SEM) in the BLA 90 min after FC.ELS animals had a decreased percentage with respect to controls.F. Representative pictures depleting DAPI, endogenous Arc::dVenus signal enhanced by GFP and c-Fos, as well as colocalization between Arc::dVenus and c-Fos (Merge).Scale bar = 100 μm.N Control-Vehicle = 7, N Control-RU486 = 8, N ELS-Vehicle = 9, N ELS-RU486 = 8. *: ELS effect.* p ≤ 0.05, ** p ≤ 0.01.

Fig. 6 .
Fig. 6.ELS decreases the number of c-Fos + in the BLA after contextual retrieval.A. Barplots showing the number of Arc::dVenus + cells per mm 2 (mean ± SEM) in the BLA 90 min after reexposure to the context 24 h later.No changes were found due to ELS or RU486 treatment.B. Barplots showing the number of c-Fos + cells per mm 2 (mean ± SEM) in the BLA 90 min after reexposure to the context 24 h later.ELS decreased the c-Fos + population.C. Barplots showing the number of double labeled Arc::dVenus + and c-Fos + cells per mm 2 (mean ± SEM) in the BLA minutes after reexposure to the context 24 h later.There were no changes between conditions.D. Barplots showing the percentage of c-Fos + cells among the total Arc::dVenus + population per mm 2 (mean % ± SEM) in the BLA 90 min after reexposure to the context 24 h later.There were no changes due to ELS or RU486 treatment.E. Barplots showing the percentage of Arc::dVenus + cells among the total c-Fos + population per mm 2 (mean % ± SEM) in the BLA 90 min after reexposure to the context 24 h later.No ELS neither RU486 treatment changed these percentages.F. Representative pictures depleting DAPI, endogenous Arc::dVenus signal enhanced by GFP and c-Fos, as well as colocalization between Arc::dVenus and c-Fos (Merge).Scale bar = 100 μm.N Control-Vehicle = 8, N Control-RU486 = 13, N ELS-Vehicle = 7, N ELS-RU486 = 8. *: ELS effect.** p ≤ 0.01.
Fig.7.ELS differentially modifies the changes in the cell percentages between training and retrieval.A. Barplots showing the number of Arc::dVenus + cells per mm 2 (mean ± SEM) in the BLA 90 min after FC versus 90 min after reexposure to the context 24 h later.Retrieval increased Arc::dVenus expression.B. Barplots showing the number of c-Fos + cells per mm 2 (mean ± SEM) in the BLA 90 min after FC versus 90 min after reexposure to the context 24 h later.Retrieval decreased the number of c-Fos + cells.C. Barplots showing the number of double labeled cells per mm 2 (mean ± SEM) in the BLA 90 min after FC versus 90 min after reexposure to the context 24 h later.Expression was decreased upon retrieval.D. Barplots showing the percentage of c-Fos + cells among the total Arc::dVenus + population per mm 2 (mean % ± SEM) in the BLA 90 min after FC versus 90 min after reexposure to the context 24 h later.Retrieval reduced this percentage.E. Barplots showing the percentage of Arc:: dVenus + cells among the total c-Fos + population per mm 2 (mean % ± SEM) in the BLA 90 min after FC versus 90 min after reexposure to the context 24 h later.This percentage was increased upon retrieval specifically in the ELS group.Training: N Control-Vehicle = 7, N Control-RU486 = 8, N ELS-Vehicle = 9, N ELS- RU486 = 8.Retrieval: N Control-Vehicle = 8, N Control-RU486 = 13, N ELS-Vehicle = 7, N ELS-RU486 = 8.*: ELS effect, &: time effect, *&: ELS by time interaction effect,^: post-hoc effect.Effect p ≤ 0.05.

Table 1
ELS from PND2-9 decreased body weight gain.Table summarizing body weight gain (mean ± Standard Error of the Mean -SEM) in control and ELS mice during the developmental period.ELS reduced body weight gain during the stress period (PND 2-9), which normalized upon returning mice to control conditions-from PND 9 until adolescence age (averaged from PND 28 to PND 30).Pre-weaning values are presented as averages per litter for male pups due to the impossibility of tracking individual pups at this early age.N Control = 10, N ELS = 11.**ELS effect, p < 0.01.