Histone Lysine Crotonylation Associated Epigenetic Mechanism Dynamically Regulates Prenatal Stress Induced Anxiety-Related Behaviour in Adolescent Offspring

Epigenetic regulation generally controls the level of transcription, histone modification possibly by the action of environmental factors, i.e., social experience, social stress, and housing condition [1]. DNA methylation is a stable, inheritable, and well-studied epigenetic mark, known to regulate brain development and function [2]. In addition, other epigenetic mark such as Malonylation (Kmal), Propionylation (Kpr), 2-Hydroxyisobutyrylation (Khib), Crotonylation (Kcr), Succinylation (Ksucc), β-hydroxybutyrylation (Kbhb), Lactylation (Kla), Glutarylation (Kglu), Butyrylation (Kbu), and Benzoylation (Kbz) has been reported to converts the extracellular signal to intracellular signal [3]. Specifically, histone modifications, i.e., histone (H3) lycine (K4/9/27) acetylation and histone (H3) lycine (K4/9/27) methylation at specific brain region can alter the behaviour, learning, and memory [4, 5]. In general, expression of methyl-CpG binding protein-2 (MeCP2) and repressor element-1 silencing transcription factor (REST) involves in the regulation of histone modifications [6].

In addition, in the process of histone crotonylation p300 catalyse covalent modification [7] and sirtuin 1 (SIRT1) act as a histone decrotonylase that remove the covalent modification [8]. Further, chromodomain Y-like (CDYL) the effectors proteins and enhancer of zeste 2 polycomb repressive complex 2 subunit (EZH2) facilitate the process of crotonyl CoA binding with histones in a crotonylation-dependent manner [9]. Earlier, histone crotonylation has been known as active transcriptional regulator and control cellular functions [10]. Epigenetic regulation fine tunes the expression of neurosecretory protein VGF; its transcriptional changes alter the brain development, anxiety/depression-like behaviour learning, and memory [11, 12].

Notably, stress associated epigenetic changes has been considered as core factor for the development of behavioural disorders [1]. In animal models, social stress during pregnancy (prenatal stress, PNS) alters the epigenetic status and associated gene expression, neurodevelopment, and behaviour later in their offspring [13, 14]. On the other side, environmental enrichment (EE) naturally considered as a positive experience enough to stimulate cognitive function, physiology and behaviour [15]. EE has been known to stimulate brain function, interact with genome, improves the synaptic plasticity, and behaviour [16, 17]. Earlier, we have reported that EE reverses the stress-induced epigenetic changes and restore their normal behaviour [18]. In this study, we examine the possible mechanism of specific epigenetic mark in PNS/EE associated signalling molecules and behaviour. To test this, female pregnant rats were exposed to social defeat (SD) stress and housed in EE until their offspring attain the age of 30 days. We found that EE normalize the behaviour in offspring possibly through histone crotonylation.

Timed female pregnant rats (Rattus norvegicus, 180–250 g) were individually housed in standard laboratory cage (43.5 × 29.0 × 15.0 cm) at animal facility (24 ± 2°C, light and dark cycle-12:12 h) with ad libitum of water and food. Female pregnant rats were grouped into: (i) control (CON; n = 4); (ii) PNS (n = 4); (iii) PNS with EE (PNS+EE; n = 4). Two housing conditions were prepared, CON and PNS animals were housed in standard laboratory condition, while PNS+EE animals housed in EE cage from gestational day (GD)-10 to until their pups’ age postnatal day (PND) −30 (Fig. 1). Male pups were selected from the dams (maximum four pups selected from one dam) that give birth an equal number of pups. Animals are not active (licking/grooming) during handling prior to the behavioural test and differences in body weight (not loosing body weight gradually or abruptly) were not considered. Animals from each group eight animals from each group (control [CON; n = 8]; PNS [n = 8]; PNS with EE [PNS+EE; n = 8]) were subjected to further behavioural analysis.

Female pregnant rats were individually subjected to SD (GD-16 to 18) to induce PNS. SD cage was constructed with two equal size (30 × 30 cm) chambers (resident chamber [RC]; intruder chamber [IC]). RC was connected with another laboratory cage to provide food and water. The wire mesh sliding door between RC and IC facilitate tactile stimulation between intruder (pregnant rat) and resident aggressor (senescence male rat). Resident animal was housed at RC (10 days) for habituation. Intruder was allowed to explore the IC (5 min) to assess the threat, and then the sliding door was opened to facilitate the physical interaction (10 min/day) with resident animal. If the intruder displayed submissive behaviour (lying on its back/showing hunchback posture) for 5 s the interaction was terminated, then the intruder was transferred to the home cage. In the absence of aggressor, control animals were allowed to explore the apparatus [19].

EE (120 × 100 × 60 cm) cage was connected with standard laboratory cage through a plastic pipe to provide ad libitum of food and water [18]. Pregnant rat was housed in EE cage from GD-10, different size, colours and shapes of objects, ladders, running wheel, plastic pipes with bends were placed in the EE cage to induce sensory, motor, cognitive, and social stimulation. Nesting materials were placed to set their own nest and suppress the aggressive behaviour among them [20]. We have observed pups with their mother in standard laboratory cage (food cage) from PND-14; however, we have not measured number of entries, time spent in the EE or standard laboratory cage.

The apparatus (100 × 100 cm white floor arena with 25 equal squares) was placed under bright light (550 lux) and test was performed for 5 min. On PND-31, animal was placed facing the centre at the corner of the apparatus and their behaviour was recorded. The number of entries into the centre square, squares crossed, and time spent in the centre square was taken for analysis.

The light and a dark chamber (20 × 40 × 20 cm each) of the apparatus connected with a guillotine door (12 × 12 cm), and constructed using plexiglass. The floor of the chambers was fixed with stainless steel (3 mm) grid (1 cm), and grid was fixed at the floor of the chambers, and the dark chamber grid alone connected with shock generator. On PND-32, animals were allowed to explore the apparatus for 5 min by placing facing away from the door at the light compartment. On PND-33, training was provided (for every 2 min) by placed animal in the light compartment and allowed to enter the dark compartment. Shock was delivered (0.5 mA, 5 s) once the animal entered the dark compartment and transferred to their home cage after 20 s. Training session was terminated if the animal stays in the light compartment continuously for 120 s. On PND-34, the same procedure was repeated to testing their memory. If the animal was not responded for 300 s, the test was terminated. Experiments were video recorded and analysed by researcher; researcher was blind to which group each animal belonged. Behavioural indices including time spent at light/dark compartment and time taken to enter compartments were calculated.

Immediately after the behavioural test (within 45 min) animals (n = 6/each group) were euthanized [sodium pentobarbital (45 mg/kg) intraperitoneally]. Rapidly, the whole brain was removed and placed on ice chilled glass plate (Petri dish) and then hippocampus (dorsal and ventral) was dissected out as described by Zapala et al. [21]. Obtained tissue was homogenized in ice-cold lysis buffer (200 μL) with protease inhibitor. Homogenate was incubated on ice (30 min) and then centrifuged (10,000 g, 4°C, 30 min). The supernatant was carefully collected and again centrifuged (12,000 g, 4°C, 15 min) to collect supernatant.

Hippocampus tissue was homogenized in TX buffer (Tris-HCl, NaCl, EDTA, and Triton-100) with cocktail of protease inhibitor. The samples were incubated on ice (15 min) and centrifuged (6,000 g, 4°C, 10 min). The supernatant was discarded, and the pellet was dissolved in HCl (0.2 m) -TX buffer. The mixture was incubated on ice (30 min) and then centrifuged (6,000 g, 4°C, 10 min) and supernatant was collected. Samples were quantified and stored at −80°C for further analysis.

The samples (60 µg) were resolved on a polyacrylamide gel (10% or 8% for higher M.wt), and then transferred onto PVDF membrane. The membranes were pre-blocked in Tris-buffered saline containing non-fat milk (5%) for 2 h at room temperature and incubated (4°C) with any one of the primary antibodies (online suppl. Table 1; for all online suppl. material, see https://doi.org/10.1159/000543696) for 12–15 h. The membrane was washed then incubated in secondary antibody conjugated with alkaline phosphatase (ALP) for 4 h. Membrane was washed with TBS and the ALP activity was detected with 5 bromo-4-chloro-3-indolyl phosphate di-sodium salt (BCIP)/nitro blue tetrazolium chloride (NBT) according to the manufacturer’s instruction. Images were obtained and band intensity was quantified using Image Lab 2 software (Molecular Imager XRS, Bio-Rad laboratories Inc. USA) and uncropped images provided in online supplementary information. Estimated intensity of the specific band values were normalized with β-actin or total-H3 accordingly.

The data were tested for statistical significance using one-way analysis of variance (ANOVA; Sigma Stat; Version 11.0) following with Bonferroni post hoc test. Graphical analysis was represented using Graphpad Prism (version 7.0).

The open field test (OFT) showed that number of squares crossed (F_(2,31) = 37.075, p < 0.001), entries to central square (F_(2,31) = 24.510, p < 0.001) and time spent at central square (F_(2,31) = 39.454, p < 0.001) was significantly varied. Post hoc analysis reported that EE significantly reduced the PNS induced anxiety-like behaviour, thus, PNS+EE (p < 0.001) and CON (p < 0.001) animals significantly crossed more squares than the PNS animals. In comparison, CON animals crossed more squares than PNS+EE animals (p < 0.01) (Fig. 2a). Further, CON (p < 0.001) and PNS+EE (p < 0.001) animals made significantly more entries to central square and spent more time than PNS animals. Whereas, we do not observed difference between CON and PNS+EE animals (p > 0.05) (Fig. 2b, c).

As shown in Figure 3, we have not observed any difference between groups (F_(2,31) = 2.10, p > 0.05) on day one during exploration, whereas, significant difference was observed in animals take time to enter the dark compartment (F_(2,31) = 95.545, p < 0.001). Comparatively, PNS animals’ latency to enter the dark compartment was significantly less (p < 0.001) than CON/PNS+EE animals, but no difference between CON and PNS+EE animals (p = 1.000) (Fig. 3a). There was a difference in the number of trials required to avoid foot shock (F_(2,31) = 21.95, p < 0.01) as on day two. PNS animals required significantly more number of trials than CON (p < 0.01) and PNS+EE (p < 0.05), but no difference between CON and PNS+EE (p > 0.05) (Fig. 3b). Their exploration time at light compartment during exploration (F_(2,31) = 132.47, p < 0.001) and test (F_(2,31) = 296.102, p < 0.001) was significantly between groups. Interestingly, CON (p < 0.001) and PNS +EE (p < 0.001) animals spent significantly more time in light compartment than PNS animals, however, detected difference between CON and PNS +EE animals was not significant (p > 0.05) (Fig. 3c). Further, there was no difference between groups in time spent at dark compartment during exploration (F_(2,31) = 1.07, p > 0.05), but significant difference was observed during testing (F_(2,31) = 143.27, p < 0.001). PNS animals spent significantly more time in the dark compartment than CON (p < 0.001) and PNS+EE (p < 0.001) animals, but no difference between CON and PNS+EE animals (p > 0.05) (Fig. 3d). Observed behavioural data demonstrate the PNS induced anxiety-like behaviour and memory impairment in their offspring, but housing at EE resilience the effect.

Further analysis (Figure 4a, online suppl. Fig. 1) showed significant differences in the level of MeCP2 (F_(2,17) = 24.039, p < 0.001) and REST (F_(2,17) = 30.937, p < 0.001) expression among experimental groups. Post hoc analysis indicated that PNS significantly induced the expression of MeCP2 and REST, level of MeCP2 was significantly higher in PNS animals than CON (p < 0.05) and PNS+EE (p < 0.05) animals. Whereas, EE significantly reduced the expression of MeCP2 and REST in PNS+EE animals, thus, there was no significant difference was detected between CON and PNS+EE animals (p > 0.05) (Fig. 4b). Similarly, the level of REST significantly induced by PNS than CON (p < 0.01) and PNS+EE (p < 0.001) animals. Whereas, EE significantly reduced the expression of PNS-induced expression of REST in PNS+EE animals, therefore, significant difference was detected between CON and PNS+EE animals (p > 0.05) (Fig. 4c). We have examined Hsp90 as a additional control marker, but there was no change in in the expression. We observed that PNS induced the expression of MeCP2 and REST in the offspring, but the EE resilience the PNS effect and suppressed the expression.

Expression of p300 (F_(2,17) = 60.246; p < 0.001) and SIRT1(F_(2,17) = 43.673; p < 0.001) was differently regulated by PNS/EE (Figure 5a, online suppl. Fig. 2). PNS suppress the expression of p300, thus the level was lower than CON (p < 0.01) and PNS+EE (p < 0.001) animals. EE reverse the PNS induced suppression and the level was higher in PNS+EE animals than CON (p < 0.01) (Fig. 5b). In contrast, SIRT1 expression was significantly induced by PNS was and higher than CON (p < 0.001) and PNS+EE (p < 0.001) animals, but no variation between CON and PNS+EE (p > 0.05) (Fig. 5c). The above results indicate that EE differently regulate the PNS induced changes in p300 and SIRT1 expression.

Expression of CDYL (F_(2,17) = 363.241; p < 0.001) and EZH2 (F_(2,17) = 64.535; p < 0.001) was significantly different between group (Figure 6a, online suppl. Fig. 3). PNS significantly induced the expression of CDYL (p < 0.001) and EZH2 (p < 0.001), which is higher than CON. In comparison, EE significantly suppress the PNS induced expression, level of CDYL (p < 0.01) and EZH2 (p < 0.001) was lower in PNS+EE animals than PNS animals. EE suppress the PNS induced expression of CYDL and EZH2; however, the level of CYDL was significantly higher than CON (p < 0.001), but no difference was observed in EZH2 (p > 0.05) (Fig. 6b, c). Observed results suggest that EE reduced the PNS induced expression of CDYL and its corepressor EZH2 expression level.

We found that PNS significantly induces triple methylation at histone 3-lycine27 (H3K27me3) (F_(2,17) = 52.972; p < 0.001) and suppress the crotonylation at histone 3-lycine18 (H3K18Cr) (F_(2,17) = 85.184; p < 0.001) (Figure 6a, online suppl. Fig. 3). Post hoc analysis showed that H3K27me3 in PNS animals was higher compared to CON (p < 0.001) and PNS+EE (p < 0.01), whereas no difference observed between the CON and PNS+EE (p > 0.05) animals (Fig. 6d). In contrast, PNS suppress the H3K18Cr in their offspring, which is significantly lower than CON (p < 0.001) and PNS+EE (p < 0.001). In comparison, no difference observed between CON and PNS+EE (p > 0.05) (Fig. 6e). Further, we observed differences in VGF expression (F_(2,17) = 23.431; p < 0.001). PNS significantly suppress the expression of VGF, which is significantly less than CON (p < 0.05) and PNS+EE (p < 0.01), whereas EE induced VGF expression was significantly higher in PNS+EE than CON (p < 0.01) (Fig. 6a, b).

Earlier, we have reported that PNS induces oxidative stress, telomere shortening, altered synaptic plasticity and memory impairment [19, 22]. In this study, we found anxiety-like behaviour and memory deficit possibly induced by PNS, which is in line with earlier reports [23, 24]. Notably, EE resilience the PNS induced effect, thus, PNS+EE offspring spent more time in the centre square and crossed more number of squares in OFT [25, 26] and spend extended duration at light compartment, reluctant to enter the dark compartment [15, 27]. Observed behavioural phenotype resembles other stress models and provide additional evidence to the beneficial effect of EE.

Earlier studies demonstrated the correlation between social stress with epigenetic changes and behaviour. PNS has been known to induce cellular and molecular changes including DNA methylation [28]. MeCP2 and REST are part in the process of methylation and have been implicated in regulation of behaviour [29, 30]. In addition, EE suppress the PNS induced expression of MeCP2 and REST [6]. MeCP2/REST play a multifunctional role during gestation, neonatal and postnatal development through transcriptional and epigenetic regulation [31, 32]. Observed behaviour in PNS/EE animals could be linked with transcriptional changes in MeCP2/REST [30, 31, 33]. Hsp90 has been known to regulate several biological mechanism beyond protein aggregation, trafficking, other cellular signalling, and metabolism [34]. Thus, we examined the expression pattern of Hsp90 to exclude possible influence of other environmental factors, immune regulation, and cellular mechanisms in our analysis. The level of Hsp90 expression was not altered, probably the basal level of Hsp90 expression within the capacity of buffer range to maintain the homeostasis [35], and other external stressful factors may not influence [36]. We have observed that EE reversed the PNS mediated suppression of p300 and induction of SIRT1 expression, which indicate that EE differently regulated the p300, SIRT1 expression. In this process, p300 and SIRT1 actively participate in epigenetic changes specific residue in the histone and regulates transcription [37‒39].

Interestingly, p300 has the histone crotonyl transferase activity and regulates transcription [3, 40]. Supporting to this, we have observed concurrence induction of CDYL expression in PNS, which is suppressed by EE. CDYL act as a negative regulator in histone crotonylation [41, 42]. CDYL reported to promote transcriptional silencing by recruiting the enhancer of zeste homolog 2 (EZH2), which regulates trimethylation of lysine 27 in histone 3 (H3K27me3) [43, 44]. Upregulated CDYL and its corepressor EZH2 in PNS animals possibly promote the H3K27me3 to repress the histone crotonylation (H3K18Cr) [9, 42]. Interestingly, EE reverse the PNS effect and normalized the level of CDYL, EZH2, H3K27me3, and H3K18Cr. The effect of PNS possibly controls the VGF promoter by H3K18Cr and H3K27me3 [42], reduced expression of VGF possibly inhibits neurogenesis, induced anxiety, and memory deficit [11, 43‒45], EE reduces the PNS induced effect and restore the normal behaviour and memory.

Taken together, this study demonstrated that environmental stimulus (PNS/EE) alters the balancing act of MeCP2, REST, p300, SIRT, CDYL, and EZH2 which alter the epigenetic status of H3K27me3 and H2K18Cr. Subsequently, regulate the expression of VGF and behaviour. Our observation suggests that PNS/EE mediated behavioural changes are possibly regulated by H2K18Cr associated epigenetic mechanism.

Experimental procedures were reviewed and approved by Institutional Animal Ethical Committee (Ref. No. BDU/IAEC/Re02/2021 dt 4/09/2021) following guidelines established by the Committee for the Purpose of Control and Supervision of Experiments on Animals (CPCSEA).

The authors have no conflicts of interest to declare.

K.E.R. thank Rashtriya Uchchatar Shiksha Abhiyan (RUSA) 2.0 – Biological Sciences and DST-FIST to Department of Animal Science. K.S. thank Indian Council for Medical Research (ICMR) for supporting through Senior Research Fellow (SRF).

K.S: investigation, validation, data analysis, writing – original draft preparation. K.E.R: conceptualization, project administration, supervision, and writing – review and editing.

All data reported in the present study. Further enquiries can be directed to the corresponding author.