Systemic HDAC Inhibition Ameliorates the Aberrant Responses to Acute Stress in Socially Isolated Male Mice

Abstract

Adverse experiences in early life can induce maladaptive responses to acute stress in later life. Chronic social isolation during adolescence is an early life adversity that can precipitate stress-related psychiatric disorders. We found that male mice after 8-week adolescent social isolation (SI) have markedly increased aggression after being exposed to a 2-hr restraint stress (RS), which was accompanied by a significant increase of AMPAR- and NMDAR-mediated synaptic transmission in prefrontal cortex (PFC) pyramidal neurons of SI^RS males. Compared to group-housed counterparts, SI^RS males exhibited a significantly decreased level of histone H3 acetylation in PFC. Systemic administrations of class I histone deacetylase (HDAC) inhibitors, romidepsin or MS-275, ameliorated the aggressive behavior, as well as general social interaction deficits, of SI^RS males. Electrophysiological recordings also found the normalization of PFC glutamatergic currents by romidepsin treatment of SI^RS male mice. These results have revealed an epigenetic mechanism and intervention avenue for aggression induced by chronic social isolation.

Graphical Abstract

Schematic diagram showing a potential mechanism underlying the aberrant response to acute stress in socially isolated male mice. Group-housed or single-housed (8-weeks, starting at 3-week-old) mice were exposed to an acute (2-hour) restraint stress, followed by behavioral assays. Chronically isolated mice exhibited dramatically heightened aggression in the Resident-Intruder (RI) test, which was accompanied by the decreased histone acetylation and increased synaptic currents mediated by NMDA and AMPA receptors in pyramidal neurons of prefrontal cortex (PFC). Treatment with class I histone deacetylase (HDAC) inhibitors, Romidepsin or MS-275, normalized PFC glutamatergic currents, and ameliorated the aggressive behavior of isolation-reared male mice.

Introduction

Social interactions play a causal role in determining health and survival in humans and other animals because of their direct impact on physiology, disease risk, and life span (Kendler et al., 1999; Snyder-Mackler et al., 2020). Chronic social isolation during adolescence is recognized as an early life adversity that can precipitate stress-related psychiatric disorders (Almeida et al., 2021; Hawkley and Cacioppo, 2010). Emotional abnormalities resulting from early social neglect include anxiety, depression and excessive aggression (Haller et al., 2014; Tan et al., 2021; Wang et al., 2022b; Zelikowsky et al., 2018). Adverse experiences in early life can also induce maladaptive responses to acute stress in later life (Holz et al., 2023; Nemeroff, 2016).

The complex interplay between genetic and environmental factors plays a crucial role in stress-induced mental disturbances (Barnett Burns et al., 2018; Rahman and McGowan, 2022; Waltes et al., 2016). One mechanism that has come into focus for mediating this interplay is epigenetics, which refers to a collection of gene regulatory processes through histone modification or DNA methylation to dynamically regulate chromatin structure and transcriptional accessibility (Strahl and Allis, 2000; Wu and Zhang, 2014). Epigenetic modifications induced by early life stress can shape the molecular response of a cell to its environment as a function of genetic predisposition (Klengel and Binder, 2015). A persistent histone acetylation increase has been revealed in the nucleus accumbens of mice exposed to chronic social defeat stress (Covington et al., 2009). In two genetically distinct mouse strains, the susceptibility or adaptation to chronic stress is determined by differential epigenetic changes in the Gdnf promoter to either repress or activate Gdnf transcription (Uchida et al., 2011). Adult male mice exposed to social isolation stress exhibit the increased global DNA methylation and enhanced activity of DNA methyltransferase in the midbrain, as well as activation of histone acetyltransferases and histone deacetylases (Hdac1 and Hdac3) and alteration of histone H3 lysine 9 acetylation (Siuda et al., 2014). These stress-induced epigenetic changes may mediate neuronal positive adaptation or negative maladaptation.

In this study, we sought to elucidate the altered behavioral response to acute stress in adult male mice after adolescent social isolation, as well as the underlying physiological alteration in key brain regions controlling emotional processes, such as the prefrontal cortex (Yan and Rein, 2022). Moreover, we examined the therapeutic potential of targeting epigenetic enzymes for behavioral and physiological aberrations induced by adolescent social isolation.

Materials and Methods

Animal and regents

All experiments were performed with the approval of the Institutional Animal Care and Use Committee (IACUC) of the State University of New York at Buffalo (PROTO202000049). C57BL/6J mice were bred and maintained in our institutional animal facility under controlled environmental conditions (22°C, 12-hour light/dark cycle) with free access to food and water. Romidepsin (Selleck Chemicals), UNC0642 (Tocris) and MS-275 (Selleck Chemicals) were dissolved in DMSO and then diluted with 0.9% saline before use. DMSO concentration of the working solution was < 0.5%. Mice were treated systemically (i.p.) with Romidepsin (1 mg/kg), MS-275 (5 mg/kg), UNC0642 (1 mg/kg) or vehicle (saline containing 0.5% DMSO) once daily for 3 consecutive days.

Social Isolation and Acute Stress Protocol

Similar to what was previously described (Chang et al., 2015; Tan et al., 2021; Wang et al., 2022b), post-weaning social isolation (SI) was carried out from postnatal day 21 (P21) to P77 (8 weeks). The mice were randomly separated into group housing (GH, 3-4 mice per cage) or individual housing (one mouse per cage). Enrichment was removed for SI (Lukkes et al., 2009), and other husbandry conditions were the same as GH. All animals were housed with ad libitum food accessibility in the 12 h light–dark cycle (light: 6 a.m. - 6 p.m.; dark: 6 p.m. - 6 a.m.). For restraint stress, each mouse (P78-81) was placed in a plastic cylinder and restrained for 2 hours. The acute stress was given 24 hrs before behavioral or electrophysiological measurements.

Behavioral Test

Animals (P79-82) were transferred to behavioral testing room (dim light) 1-2 hours before testing. All behavioral experiments were carried out during the day (10:00 am – 3:00 pm). Animals were used to handling before testing.

Elevated plus-maze (EPM) test:

Mice were placed in the center of a plus maze that was elevated 50 cm above the floor with two opposite open arms and two opposite closed arms. (each arm was 88 cm long, 28 cm-height walls only on closed arms) arranged at right angles. The number of entries and time spent in the closed and open arms were monitored for 5 min.

Resident intruder (RI) Test:

A modified protocol (Chang et al., 2015) was used. Briefly, the group-housed mice were individually housed for one day prior to the RI test, while the home cages of single-housed mice were not changed. In RI test, the “resident” mouse was exposed to an “intruder”, a slightly smaller (5%–15% lighter) unfamiliar control mouse of the same sex, in the home cage for 10 min. The intruders were all socially experienced, but naive to the RI test. Each intruder was used only once and was not re-used for other aggressive encounters (to avoid winner or loser effects). Aggressive behavior of the resident mouse against the intruder, including lateral threat, upright posture, clinch attack, keep down, and chase (Lukkes et al., 2009; Miczek et al., 2001), were scored to measure aggression level. The number and time of social interactions between the resident mouse and the intruder, as well as the locomotive activity of the resident mouse, were tracked using a custom software written on top of OpenCV in Python (Bradski, 2000). Non-aggressive interactions were calculated by subtracting aggressive events from total interactions.

Immunocytochemical Staining

Animals were anesthetized with a mixture of ketamine/xylazine (100/10 mg/kg) before perfusion. After perfusion with 4% fresh prepared paraformaldehyde (PFA), brains were collected, post-fixed in 4% PFA overnight, and dehydrated in 30% sucrose for 2 days. Brain slices (50 μm) containing mouse PFC, which is comprised of anterior cingulate cortex (ACC), prelimbic (PL), and infralimbic (IL), were cut coronally for staining. After washing with PBS for 3 times (10 min each), slices were blocked in PBS containing 5% BSA and 0.3% Triton for 2 hrs at room temperature. Then slices were incubated with primary antibody against H3K9Ac (1:1000, Cell signaling technology, 9649) and NeuN (1:500, Millipore Sigma, MAB 377) overnight at 4 °C. After washing with PBS for 3 times (15 min each), slices were incubated with secondary antibodies, Alexa Fluor 488-anti rabbit (1:1000; Invitrogen, A-11008) and Alexa Fluor 568-anti mouse (1:1000; Invitrogen, A-11004), for 2 hrs at room temperature. DAPI (ThermoFisher, D1306, 5 μg/ml) was used to stain nucleus for 30 min at room temperature. ProLong^™ Diamond Antifade Mountant (Invitrogen, P36970) was used to prepare microscope slides. Images were acquired using 63× objective on a Leica TCS SP8 confocal microscope. All specimens were imaged under identical conditions and analyzed with identical parameters using Image J software (version 1.52p, NIH).

Electrophysiological Recording

Mouse was decapitated under 1-3% isoflurane (Sigma) anesthesia and the brain was quickly removed and coronally cut into 300-μm slices with a vibratome (Leica VP1000S, Leica Microsystems Inc.) in an ice-cold sucrose solution. The slices were recovered and maintained at room temperature (22°C) in standard artificial cerebrospinal fluid (ACSF; in mM: 130 NaCl, 26 NaHCO₃, 3 KCL, 5 MgCl₂, 1.25 NaH₂PO₄, 1 CaCl₂, 10 Glucose) for at least 1 hr. The slice was transferred into a recording chamber on an upright microscope (Olympus) and perfused with oxygenated ACSF. Neurons were viewed under a water-immersion lens (40×) and a CCD camera. A Multiclamp 700 A amplifier with Clampex 8.2 software and Digidata 1322A (Molecular Devices, Sunnyvale, CA) were used for recordings. A pipette puller (Model P-97, Sutter Instrument Co.) was used to pull recording pipettes from glass capillaries (1.5 mm OD and 0.86 mm ID) with resistance at 3-4 MΩ.

Whole-cell voltage-clamp recording was used to measure synaptic currents in mouse PFC (ACC and PL) layer V pyramidal neurons (Tan et al., 2019; Wang et al., 2018). Pipette was filled with the intracellular solution (in mM: 130 Cs-methanesulfonate, 10 CsCl, 4 NaCl, 1 MgCl₂, 10 HEPES, 5 EGTA, 2 QX-314, 12 phosphocreatine, 5 MgATP, 0.5 Na3GTP, 0.1 leupeptin, pH 7.3, 270 mOsm). A stimulating electrode (FHC, Bowdoinham, ME) was placed ~100 μm away from the recorded neuron. Excitatory postsynaptic currents (EPSC) were elicited by a series of pulses from an S48 stimulator (Grass Technologies, West Warwick, RI) with different intensities that delivered at 0.05 Hz. For input–output responses, EPSC was evoked by a series of pulses with different stimulation intensities (1–7 V) delivered at 0.05 Hz. For AMPAR-EPSC, the membrane potential was maintained at −70 mV. For NMDAR-EPSC, the cell (clamped at −70 mV) was depolarized to +40 mV for 3 s before stimulation to fully relieve voltage-dependent Mg²⁺ block. The peak of NMDAR-EPSC was calculated at 40 ms after the onset of mixed EPSC when AMPAR was inactivated (Zhong et al., 2022). GABA_AR-IPSC was recorded with a holding potential of 0 mV using the same internal solution. Spontaneous EPSCs (sEPSC) and IPSC (sIPSC) were recorded with the same external and internal solutions in neurons clamped at −70 mV and 0 mV, respectively.

Statistics

A custom-made python script based on Neo (Garcia et al., 2014) was used for electrophysiological data analyses. Statistical analysis of the data was performed using the rstatix framework (Kassambara, 2022) for the R statistical software. All data were presented as means ± SEM. Experiments with two groups were analyzed statistically using unpaired Student’s t-tests with Welch’s correction, unless the data failed Shapiro-Wilk tests for normality, in which case the data were subjected to Mann-Whitney U (M-W) test. Experiments with more than two groups were subjected to an ANOVA (one-way or repeated measure two-way) or a Kruskal-Wallis test, followed by post hoc comparisons with Bonferroni corrections for multiple comparisons.

Results

Chronic social isolation leads to an aberrant behavioral response after acute stress.

To find out the impact of chronic social isolation (SI) on the response to acute stress, we gave a 2-hr restraint stressor (RS) to male mice either single housed for 8 weeks (starting at 3 weeks old) or group housed (GH) throughout the rearing. First, we carried out resident intruder (RI) test, a paradigm with high face and construct validity for aggressive behaviors (Koolhaas et al., 2013). As shown in Figure 1A, compared to GH^RS control mice, SI^RS mice exhibited a small but significant decrease in the number of total interactions, but a dramatic increase in the number and duration of aggressive interactions against intruders, suggesting a maladaptive response to acute stress, particularly the heightened aggression, consistent with previous findings (Chang et al., 2018; Nordman et al., 2020). Interestingly, the number of non-aggressive interactions was also significantly reduced in SI^RS mice, suggesting the impairment of general social interaction. During RI tests, the locomotion of testing mice did not reveal significant differences between groups (Figure 1B), suggesting that the aggressive behavior of SI^RS mice was not due to locomotive abnormality.

Figure 1. Socially isolated male mice exposed to acute stress has escalated aggression.

(A) Bar graphs showing aggressive behavior in the Resident-Intruder (RI) test of group-housed and single-housed socially isolated mice after an acute restraint stress (GH^RS and SI^RS, n = 9-10 mice/group; # Total interactions: t₁₁ = 2.3, p = 0.04; # Aggressive interactions: t₉ = 5.2, p = 0.0006; # Non-aggressive interactions: t₁₀ = 5.2, p = 0.0004; Total interaction time: t₁₄ = 1.7, p = 0.1; Aggressive interaction time: t₉ = 4.2, p = 0.002; Non-aggressive interaction time: t₁₆ = 1.1, p = 0.3, t-test). (B) Bar graphs of the overall distance traveled by each group while performing the RI test (n = 9 mice/group; t₁₀ = 0.1; p = 0.9, t-test). (C) Bar graphs of time, distance traveled and number of entries to Open arms (OA) and Closed arms (CA) for both groups in the elevated plus maze (EPM) test (n = 10 mice/group; OA time: t₁₇ = 0.3, p = 0.8; CA time: t₁₇ = 0.6, p = 0.6; OA distance: t₁₅ = 0.8, p = 0.4; CA distance: t₁₈ = −2.8, p = 0.01; OA entries: t₁₆ = 0.2, p = 0.8; CA entries: t₁₈ = 2.0, p = 0.07, t-test). In all figures, *: p < 0.05, **: p < 0.01, ***: p < 0.001. All data are expressed as mean ± SEM.

Next, both groups were tested in the Elevated Plus Maze (EPM), a widely used behavioral test to assess anxiety-related behaviors (Walf and Frye, 2007). As shown in Figure 1C, SI^RS and GH^RS mice exhibited no significant differences on most parameters, including time in open arm, open arm entry numbers, and distance traveled on open arm, suggesting that SI^RS mice do not have apparent anxiety.

Isolated mice exposed to acute stress have the increased glutamatergic transmission in PFC pyramidal neurons.

To determine what might drive the aberrant response to acute stress in SI^RS mice, we examined the synaptic activity in PFC, a brain region playing a key role in regulating stress-induced excessive aggressive behavior (Nelson and Trainor, 2007; Zhang et al., 2022). Whole-cell patch-clamp recordings were performed to measure spontaneous excitatory postsynaptic current (sEPSC) and AMPAR- or NMDAR-mediated EPSC evoked by electrical stimulation, as well as spontaneous and evoked GABA_AR-mediated inhibitory currents (IPSC) in layer V PFC pyramidal neurons. Evoked synaptic currents reflect the integrated response from simultaneous activation of multiple synapses that have action potential-driven synchronous or asynchronous release of vesicles. Spontanous synaptic currents reflect the response induced by randomly activated synapses that have action potential-independent release of single synaptic vesicles.

As shown in Figure 2A and 2B, the amplitudes of NMDAR- and AMPAR-EPSC evoked by a series of stimulation intensities were significantly larger in SI^RS mice, compared to GH^RS mice. The input/output curve of GABA_AR-IPSC was similar in SI^RS vs. GH^RS mice (Figure 2C). The amplitude or frequency of sEPSC or sIPSC was also not significantly different between the two groups (Figure 2D, 2E). Taken together, these results indicate that the aberrant behavioral response to acute stress in SI^RS mice is accompanied with an increase of AMPAR- and NMDAR-mediated synaptic transmission in PFC pyramidal neurons.

Figure 2. Socially isolated male mice exposed to acute stress show increased glutamatergic transmission in PFC pyramidal neurons.

(A, B, C) Input-output curves and representative traces of NMDAR-EPSC (A), AMPAR-EPSC (B) and GABA_AR-IPSC (C) in PFC pyramidal neurons from GH^RS and SI^RS mice (n = 12-19 cells from 12-19 slices in 7 mice/group, NMDAR-EPSC: F_{1,23 (group)} = 12.9, p = 0.002; AMPAR-EPSC: F_{1,31 (group)} = 10.3, p = 0.003; GABA_AR-IPSC: F_{1,33 (group)} = 1.8, p = 0.2, two-way rmANOVA). (D, E) Bar graphs and representative traces of spontaneous EPSC (sEPSC, D) and IPSC (sIPSC, E) in PFC pyramidal neurons from GH^RS and SI^RS mice (n = 15-17 cells from 15-17 slices in 7 mice/group; sEPSC: Amplitude: t₃₀ = 0.6, p = 0.6; Frequency: t₃₀ = 0.6, p = 0.6, t-test; sIPSC: Amplitude: t₃₀ = 1.1, p = 0.3; Frequency: t₃₀ = 0.6, p = 0.6, t-test). In all figures, *: p < 0.05, **: p < 0.01, ***: p < 0.001. All data are expressed as mean ± SEM.

HDAC inhibition rescues the aberrant behavioral response to acute stress in socially isolated male mice.

Since the activity of class I HDACs is found to be upregulated after various forms of chronic stress (Covington et al., 2009; Siuda et al., 2014; Uchida et al., 2011), we also compared the level of histone acetylation in deep and intermediate layers of PFC subregions from group-housed vs. socially-isolated male mice after acute stress.. As shown in Figure 3A and 3B, the fluorescent signal of H3K9 acetylation (H3K9Ac) were significantly decreased in layer 5/6 ACC, layer 2/3 ACC and layer 2/3 PL neurons of SI^RS mice, compared with GH^RS mice. No significant changes of H3K9Ac were found in layer 5/6 IL & PL or layer 2/3 IL. When all data were pooled together, the fluorescent signal of H3K9Ac in layer 5/6 and layer 2/3 medial PFC neurons of SI^RS mice were significantly lower than those from GH^RS mice.

Figure 3. Histone acetylation is reduced in PFC of socially isolated male mice exposed to acute stress.

(A) Representative confocal images of immunostaining of H3K9Ac (green) and NeuN (red) in different layers of PFC subregions, including ACC-L5/6, ACC-L2/3, PL-L5/6, PL-L2/3, IL-L5/6 and IL-L2/3, of GH^RS and SI^RS mice. Scale Bars: 10 μm. (B) Quantification of H3K9Ac fluorescent intensity (normalized with NeuN) in different layers of PFC subregions of GH^RS and SI^RS mice (n = 6 images from 6 slices in 3 mice/group). ACC-L5/6, t₁₀ = 2.4, p = 0.04; ACC-L2/3, t₁₀ = 2.3, p = 0.04; PL-L2/3, t₁₀ = 3.3, p = 0.008; mPFC-L5/6, t₁₀ = 2.7, p = 0.01; mPFC-L2/3, t₁₀ = 2.4, p = 0.02; unpaired t test. In all figures, *: p < 0.05. All data are expressed as mean ± SEM.

Next, we treated SI^RS mice with a highly potent and brain permeable class I HDAC inhibitor Romidepsin (0.25 mg/kg, i.p. 3×) (Qin et al., 2018; Zhang et al., 2021). Another structurally different class I HDAC inhibitor, MS-275 (5 mg/kg, i.p., 3×) (Ma et al., 2018; Wang et al., 2022a), was also used to confirm the involvement of HDAC. Two hours after the 3-day treatment, animals were given an acute stressor, followed by behavioral testing the next day. As shown in Figure 4A, compared to vehicle-treated controls (SI^RS+Veh), SI^RS mice treated with Romidepsin (SI^RS+Rom) exhibited a significant reduction of the number and time of aggressive interactions in the RI test, and SI^RS mice treated with MS-275 (SI^RS+MS275) also had significantly reduced aggressive interaction time. Interestingly, the reduced non-aggressive interactions in SI^RS mice were improved by Romidepsin or MS-275 treatment, suggesting their pro-social effects.

Figure 4. Treatment with class I HDAC inhibitors ameliorates hyper-aggressive behavior and general social interaction deficits in socially isolated male mice exposed to acute stress.

(A) Bar graphs showing aggressive behavior in the Resident-Intruder (RI) test of GH^RS and SI^RS mice treated with vehicle, romidepsin, UNC0642 or MS-275 (n = 7-12 mice/group; # Total interactions: F_5,53 = 0.8, p = 0.6; # Aggressive interactions: F_5,53 = 12.3, p < 0.0001; # Non-aggressive interactions: F_5,53 = 3.6, p = 0.007; Total interaction time: F_5,53 = 0.9, p = 0.5; Aggressive interaction time: F_5,53 = 9.1, p < 0.0001; Non-aggressive interaction time: F_5,53 = 3.5, p = 0.009, one-way ANOVA). (B) Bar graph of the overall distance traveled while performing the RI test (n = 8-13 mice/group, F_5,55 = 2.27, p = 0.06, one-way ANOVA). (C) Bar graphs of time, distance traveled and number of entries to Open arms (OA) and Closed arms (CA) for each group in the elevated plus maze (EPM) test (n = 6-12 mice/group, OA time: F_5,51 = 1.2, p = 0.3; CA time: F_5,51 = 1.0, p = 0.4; OA distance: F_5,51 = 2.2, p = 0.07; CA distance: F_5,51 = 1.0, p = 0.4; OA entries: F_5,51 = 1.7, p = 0.2; CA entries: F_5,51 = 1.1, p = 0.4, one-way ANOVA). In all figures, *: p < 0.05, **: p < 0.01, ***: p < 0.001. All data are expressed as mean ± SEM.

To find out the specificity of this therapeutic effect of HDAC inhibitors, we tested another epigenetic drug, UNC0642 (1 mg/kg, i.p., 3×) (Wang et al., 2021; Zheng et al., 2019), a highly selective and potent inhibitor of histone methyltransferase EHMT1/2. As shown in Figure 4A, treatment with UNC0642 failed to reduce the numbers or time of aggressive interactions in SI^RS mice, suggesting the lack of involvement of EHMT.

To ensure that the observed reduction of aggressive behavior was not caused by drug-induced changes in locomotion, we tracked locomotor activity of mice during RI tests. No significant differences in locomotion were observed among GH^RS or SI^RS mice treated with vehicle, Romidepsin, UNC0642 or MS-275 (Figure 4B).

In addition, we used EPM to test anxiety in SI^RS or GH^RS mice treated with various drugs. As shown in Figure 4C, Romidepsin-treated SI^RS mice exhibited a trend of increase in open arm time and distance, but none of the measured parameters were significantly changed. Collectively, these data suggest that HDAC inhibition can ameliorate the escalated aggressive behavior, as well as general social interaction deficits, in SI^RS mice.

HDAC inhibition normalizes glutamatergic transmission in socially isolated mice exposed to acute stress.

Given the therapeutic effects of HDAC inhibitors on aggressive behavior of SI^RS mice, we next examined whether treatment with Romidepsin could reverse synaptic dysfunction in PFC pyramidal neurons. As shown in Figure 5A and 5B, NMDAR-EPSC and AMPAR-EPSC amplitudes in Romidepsin-treated SI^RS mice (SI^RS+Rom) were significantly smaller than those in vehicle-treated SI^RS mice (SI^RS+Veh) and were similar to those in vehicle-treated GH^RS controls (GH^RS+Veh).

Figure 5. HDAC inhibition normalizes glutamatergic transmission in PFC pyramidal neurons from socially isolated male mice exposed to acute stress.

(A, B, C) Input-output curves and representative traces of NMDAR-EPSC (A), AMPAR-EPSC (B) and GABA_AR-IPSC (C) in PFC pyramidal neurons from GH^RS and SI^RS mice treated with vehicle or romidepsin (n = 7-11 cells from 7-11 slices in 6-7 mice/group, NMDAR-EPSC: F_{3,24 (group)} = 8.0, p = 0.0007; AMPAR-EPSC: F_{3,36 (group)} = 7.6, p = 0.0004; GABA_AR-IPSC: F_{1,35 (group)} = 1.6, p = 0.2, two-way rmANOVA). # p < 0.05, ## p < 0.01, ### p < 0.001, SI+Veh vs. GH+Veh; * p < 0.05, ** p < 0.01, *** p < 0.001, SI+Veh vs. GH+Rom; † p < 0.05, †† p < 0.01, ††† p < 0.001, SI+Veh vs. SI+Rom. (D, E) Bar graphs of amplitude and frequency of sEPSC (D) and sIPSC (E) in PFC pyramidal neurons from vehicle- or romidepsin-treated GH^RS or SI^RS mice (n = 10-11 cells from 10-11 slices in 6-7 mice/group, sEPSC: Amplitude: F_3,38 = 0.7, p = 0.6, Frequency: F_3,38 = 0.4, p = 0.7; sIPSC: Amplitude: F_3,38 = 6.0, p = 0.002, Frequency: F_3,38 = 0.3, p = 0.9, one-way ANOVA). Inset: representative traces. All data are expressed as mean ± SEM.

Romidepsin treatment did not significantly alter GABA_AR-IPSC in SI^RS mice (Figure 5C). Furthermore, both sEPSC and sIPSC remained largely unchanged after treatment with Romidepsin, despite a small but significant increase in the amplitude of sIPSC (Figure 5D and 5E). Together, these results suggest that the abnormally elevated glutamatergic transmission in SI^RS mice can be reversed by the inhibition of class I HDAC.

Discussion

In this study, we have revealed a maladaptive response to acute stress in male mice after chronic post-weaning social isolation -- the drastically increased aggression. After submitting to a 2-hr restraint stress, socially isolated mice were hyper-aroused during aggressive contacts and delivered substantially more attacks in the resident-intruder test. Consistently, human and animal data also show that adverse early-life events, such as repeated maternal separation, adolescent social isolation and peripubertal stress, strongly affect emotional responses to conflict and aggressive behavior in adulthood (Haller et al., 2014). Heightened aggression has been associated with various psychiatric disorders, including post-traumatic stress disorder (Taft et al., 2017) and paranoid schizophrenia (Darrell-Berry et al., 2016). The escalated aggression in socially isolated mice is an indicator of their elevated vulnerability to mental distress and disorders.

Neuroimaging studies suggest that impulsivity and aggression are correlated with frontal and temporal brain abnormalities (Soyka, 2011). Using the activity marker c-Fos, it has been found that post-weaning social isolation in rats submitted to resident–intruder conflicts leads to significantly increased activation of brain areas controlling inter-male aggression, such as the medial and lateral orbitofrontal cortices, anterior cingulate cortex, medial and basolateral amygdala, and hypothalamic paraventricular nucleus (Toth et al., 2012). Our electrophysiological studies have found that SI males after 2-hr restraint stress have significantly larger NMDAR- and AMPAR-mediated synaptic currents in PFC pyramidal neurons, compared to GH counterparts. The hyperactive PFC in response to acute stress could lead to dysregulation of direct and indirect targets of PFC, which contributes to the manifestation of aggressive behavior (Adams and Rosenkranz, 2016; Tan et al., 2021; Wei et al., 2018).

A potential mechanism for gene-environment interactions in shaping risks for stress-related psychiatric disorders is epigenetic modifications that can lead to prolonged molecular changes (Klengel and Binder, 2015). Mounting evidence has revealed a plethora of epigenetic changes that are induced by various types of stress, which cause aberrant DNA methylation and histone modification at specific genes in target regions including prefrontal cortex (Sanacora et al., 2022). In blood samples from schizophrenia patients who have encountered early life stress (ELS), the level of HDAC1 is increased (Bahari-Javan et al., 2017). Moreover, Hdac1 over-expression in mouse PFC neurons mimics schizophrenia-like phenotypes induced by ELS, and systemic administration of an HDAC inhibitor rescues the detrimental effects of ELS (Bahari-Javan et al., 2017). In human subjects with depression and male mice subjected to chronic social defeat stress (CSDS), RAC1, a key regulator of actin cytoskeleton and synaptic structure, shows the reduced transcription, which is associated with a repressive chromatin state surrounding its promoter (Golden et al., 2013). Inhibition of class I HDACs rescues CSDS-induced decrease of RAC1 transcription and depression-related behaviors (Golden et al., 2013). In the current study, we have found the reduced histone acetylation in PFC of SI males exposed to acute restraint stress (SI^RS), compared to GH counterparts. More importantly, treatment with class I HDAC inhibitors, romidepsin or MS-275, prevented the heightened aggression in SI^RS males, while an inhibitor for histone methylation enzymes EHMT1/2 was ineffective. These studies highlight the potential of HDAC inhibition as a therapeutic avenue for stress-associated psychiatric disorders.

To understand the physiological basis of the behavioral impact of HDAC inhibitor in SI^RS males, we examined its effect on PFC glutamatergic transmission, which mediates emotional and cognitive processes subserved by PFC (Yan and Rein, 2022). We have revealed that the abnormally elevated NMDAR-EPSC and AMPAR-EPSC in PFC pyramidal neurons of SI^RS males are reversed by romidepsin treatment, suggesting the normalization of neuronal activity in PFC and its target regions by HDAC inhibition. Taken together, this study has revealed an epigenetic mechanism that may be causally linked to aberrant responses to acute stress in male mice exposed to adolescent isolation stress.

Key Points.

• Adolescent chronic social isolation can precipitate stress-related psychiatric disorders.

• A significant increase of glutamatergic transmission is found in prefrontal cortex (PFC) of socially isolated male mice exposed to an acute stress (SI^RS).

• Treatment with class I histone deacetylase (HDAC) inhibitors ameliorates the aggressive behavior and social interaction deficits of SI^RS males, and normalizes glutamatergic currents in PFC neurons.

• It provides an epigenetic mechanism and intervention avenue for aberrant stress responses induced by chronic social isolation.

Biography

Luis Gustavo Hernandez Carballo is currently a Postdoctoral Fellow in State University of New York at Buffalo, School of Medicine and Biomedical Sciences. He earned Master’s and PhD degrees in Physiological Sciences from Benemérita Universidad Autónoma De Puebla in Mexico. His research primarily centered around the synaptic action of neuromodulators associated with mental health and illnesses. He has over ten years of experience in electrophysiological techniques and data analysis tool development. Recently, he has been investigating epigenetic abnormalities caused by genetic and environmental factors, contributing to synaptic dysfunction and behavioral deficits.

Data availability

Data will be made available on request.