Perineuronal nets, inhibitory interneurons and anxiety-related ventral hippocampal neuronal oscillations are altered by early life adversity

Abstract

BACKGROUND:

In humans, accumulated adverse experiences during childhood increase the risk of anxiety disorders and attention-deficit hyperactivity disorder. In rodents, the ventral hippocampus (vHIP) is associated with anxiety regulation, and lesion of this region alters both anxiety-like behavior and activity levels. Neuronal oscillations in the vHIP of the theta frequency range (4–12 Hz) have been implicated in anxious states and derive in part from the activity of inhibitory interneurons in the hippocampus, some of which are enwrapped with perineuronal nets (PNNs), extracellular matrix structures that regulate plasticity. Here we sought to investigate the associations among early life stress-induced anxiety and hyperactivity with vHIP neuronal oscillations, inhibitory interneurons and PNNs in mice.

METHODS:

We used repeated maternal separation with early weaning (MSEW) to model accumulated early life adversity in mouse offspring and studied the underlying behavioral and cellular level changes in the vHIP associated with early life adversity.

RESULTS:

We found increased anxiety-like behavior and activity levels in MSEW adult males, along with increased theta power and enhanced theta-gamma coupling in the vHIP. MSEW mice showed reduced intensity of parvalbumin as well as increased PNN intensity around parvalbumin+ interneurons in the vHIP. We further observed that MSEW increased OTX2, a transcription factor promoting PNN development, in the choroid plexus where it is produced, as well as in parvalbumin+ interneurons, where it is sequestered.

CONCLUSION:

These findings raise the possibility of causal links among parvalbumin+ interneurons, PNNs, OTX2, and MSEW-induced anxiety and hyperactivity.

Introduction

Childhood maltreatment is known to substantially increase the risk of developing neuropsychiatric conditions in adulthood, including anxiety and mood disorders (1–3). In fact, more than 50% of diagnosed cases of anxiety disorders and clinical depression occur in individuals who have experienced early life adversity (4). Attention-deficit hyperactivity disorder (ADHD) is often comorbid with anxiety disorders in adults and childhood maltreatment has been shown to increase susceptibility to both ADHD and excessive anxiety (5). Studies in humans suggest that exposure to stress during a sensitive period in development may be predictive of negative outcomes (6–8), but the preponderance of evidence, including a recent large study, indicates that the accumulation of multiple bouts of early life stress exposure is most damaging in terms of long-term psychiatric problems (7–10). To cover both the sensitive period and cumulative aspects of early life adversity, we selected a multiple-hit model of early life stress in mice, maternal separation with early weaning (MSEW), where mouse pups are subjected to daily maternal separations of increasing length during the first 16 days of life followed by weaning four days earlier than typically reared laboratory mice (11; 12). This developmental manipulation has been shown to increase both anxiety levels and hyperactivity in adulthood (11), suggesting it has strong translational validity (see Supplemental Materials for further discussion).

The ventral hippocampus (vHIP) of rodents has been linked to anxiety regulation; bilateral lesions of this area have been shown to reduce anxiety-like behavior in rodents (13–15) while optogenetic or pharmacological stimulation of the region has the opposite effect (16–18). Several studies suggest that neuronal oscillations in the theta frequency (4–12 Hz) range in the vHIP and mPFC are important for anxiety regulation, with decreases in theta power in the vHIP occurring after anxiolytic drug administration (19) and increases in theta power in the vHIP and mPFC as well as theta synchrony between the vHIP and mPFC during anxious states (20). Furthermore, in humans, increased self-reported anxiety during threat is associated with increased theta rhythm in the anterior hippocampus (21), a region that is considered analogous to the vHIP in rodents (22). Despite these associations, the electrophysiological profile of rodents that display increased anxious behavior after early life adversity has not been studied.

Inhibitory interneuron subtypes have been implicated in theta rhythm in the hippocampus and neocortex (23–27). Optogenetic studies on the hippocampus have shown that parvalbumin (PV)+ interneurons are the main effectors of intrinsic theta rhythm while somatostatin (SST)+ interneurons modulate the entrainment of intrinsic theta by controlling external inputs to the hippocampus (25). Many PV+ interneurons are encapsulated by specialized extracellular matrix structures termed perineuronal nets (PNNs). PNNs are thought to limit plasticity and are responsible for the closure of the critical period for ocular dominance in the visual cortex (28). PNN degradation has been shown to reduce PV+ interneuron activity (29), as well as reduce theta frequency synchrony between brain regions involved in fear memory (30). Experience has been shown to alter PNN expression but the effects are complex, depending on the type of experience, the developmental stage examined and the brain region. For example, early life trauma has been shown to reduce PNN expression in the basolateral amygdala (31), enriched environment rearing reduces PNNs in the cerebellum (32) but increases them in the CA2 region of the hippocampus (33), while enriched environment exposure in adulthood increases or decreases PNN expression depending on the brain region and presence of other experiences (34). PNNs have been linked to the transcription factor orthodenticle homeobox protein 2 (OTX2) (35), the expression of which coincides with the maturation of PV+ cells and the closure of the critical period of plasticity in mice. In adults, OTX2 is produced in the choroid plexus, released into the CSF and accumulates in PV+ cells where it facilitates PNN formation. No previous studies have investigated whether PNNs and OTX2 are altered in PV+ cells of the hippocampus after early life adversity.

Studies have shown that adult-generated neurons in the hippocampus have the ability to modulate neuronal oscillations in that their elimination by focal irradiation or by transgenic means increases rhythmic firing (36). Some previous studies have reported that early life adversity reduces the production of new neurons in the dentate gyrus (37), raising the possibility that changes in neuronal oscillations and potentially anxiety may involve reductions in this form of plasticity. Additionally, adult-generated neurons also form connections with PV+ inhibitory interneurons (38; 39). Despite these compelling associations, no previous studies have investigated whether changes in inhibitory interneurons, PNNs and adult-generated neurons occur in association with early life stress-induced alterations in neuronal oscillations and anxiety-like behavior.

Here, we sought to investigate these possibilities by performing behavioral analyses and local field potential (LFP) recordings in the vHIP of adult mice subjected to MSEW and found consistent increases in anxiety-like behavior and hyperactivity along with increases in theta power in a novel environment, as well as enhanced theta-gamma coupling in both familiar and novel environments. This was accompanied by a reduction in densities of PV+ and SST+ interneurons, but no decrease in immature neurons, in the granule cell layer (GCL) of MSEW mice. Further analysis of PV+ interneurons revealed increased PNN intensity surrounding PV+ cells as well as increases in transcription factor OTX2 within these cells. These findings suggest interrelationships among PV+ interneurons, PNNs, OTX2, and MSEW-induced anxiety and hyperactivity.

Methods and Materials

Maternal Separation with Early Weaning

Twenty litters of C57BL/6J mice from two breeding cohorts were exposed to one of two rearing conditions at P2: 1) non-handled controls; or 2) maternal separation for 4h daily from P2-P5 and 8h daily from P6-P16. Pups in the latter group were weaned at P17 while control litters were weaned normally at P21 (11). At age P60-P70, mice underwent behavior testing. See supplemental methods for more details.

Behavior

Anxiety and activity testing were carried out as detailed in supplemental methods.

Histochemistry

Two hours after behavioral analyses were conducted, mice were perfused and sections through the hippocampus were stained for doublecortin, Ki67, calretinin, PV, SST, OTX2, cfos and WFA using immunohistochemical and histochemical methods. Antibodies are detailed in Table S1. See supplemental methods for more details.

Cell count and density analyses

The vHIP was analyzed separately as it is functionally distinct from the dHIP and has been implicated in anxiety regulation (22). DCX+, Ki67+ and GAD67+ cells from the GCL and interneuron subtypes across layers of the dentate gyrus (DG), CA1 and CA3 were counted using Stereo Investigator software (Microbrightfield Bioscience). Densities were determined by dividing the total number of positive cells by the volume of the region outlined.

WFA and OTX2 intensity analyses

Confocal images were analyzed on ImageJ where every PV+/WFA+ cell was outlined and the area and maximum intensity (mean gray value of WFA or OTX2 x area of the WFA stain) for each WFA+/PV+ cell was determined according to previously published protocols (40; 41). See supplemental methods for more details.

Surgery

Insulated stainless steel electrodes (0.005”, Plastics One) were implanted in the vHIP CA1; (3.3 mm posterior, 3.45 mm lateral and 4.2 mm depth). Placement of electrode was in the left hemisphere as this hemisphere has been specifically shown to be affected in MSEW mice (42).

LFP Recordings and Analysis

One week after surgery, mice from cohort 2 underwent habituations and behavioral testing. LFP recording data were collected on the fourth day from 2 separate recording sessions. First, mice were exposed to a pre-habituated familiar environment for a period of 10 min followed by an hour of rest and then exposed to a novel environment for a period of 10 min. Power spectra were calculated from data acquired from mice during segments of movement only (4–15 cm/s). Any animal without a minimum of 5 seconds of movement during the task was excluded from analysis. For phase amplitude coupling analyses (43), all animals regardless of movement (moving or stationary) were considered for analysis. See supplemental methods for details.

Statistical analysis

Electrophysiology data were analyzed as described above using custom Matlab and R scripts (code available upon request). For all other measures, unpaired two tailed Student’s t-tests or Mann-Whitney U tests were performed on each data set following determination of homogeneity of variance with Levene’s test. All statistical tests and p values are listed in Tables S2 and S3.

Results

MSEW results in increased anxiety-like behaviors in males in the elevated plus maze

We observed increased anxiety-like behavior in 2 cohorts of MSEW male mice compared to male controls. MSEW mice in cohort 1 showed increased anxiety in the EPM with significantly lower percentage of entries into the open arms (Fig 1B). MSEW mice in cohort 1 also spent significantly more time in the closed arms (Fig 1C). To verify the reliability of this model, we repeated behavioral tests with a second cohort of mice. MSEW mice from cohort 2 also displayed greater anxiety in the EPM as they made significantly lower percent entries into the open arms (Fig 1D), although no differences were observed in the amount of time spent in the open or closed arms between groups (Fig 1E). No measurable changes in anxiety-like behavior were observed in MSEW female mice compared to same sex controls (Fig S1). Because our aim was to search for neural correlates of MSEW-increased anxiety, we focused our subsequent studies on males.

Figure 1.

(A) Timeline depicting the MSEW paradigm. (B) MSEW male mice from cohort 1 displayed increased anxiety-like behavior in the EPM, making significantly lower percent entries in to the open and higher entries into closed arms. (C) MSEW mice spent significantly more time in the closed arms of the EPM. (D) MSEW male mice from cohort 2 displayed increased anxiety-like behavior in the EPM, making significantly lower percent entries into the open arms and higher percent entries into closed arms. (E) MSEW mice from cohort 2 did not show any differences in percent time spent in the open or close arms. *p<0.05

MSEW results in increased activity levels in a novel environment

Consistent with previous work (11), we found that MSEW male mice were more active in a novel environment than controls. MSEW male mice exhibited greater locomotion in a novel testing arena in that they covered a larger distance (Fig 2A) as well as an increase in the percentage of time spent moving during the duration of test (Fig 2B).

Figure 2.

(A) MSEW male mice display greater activity in a novel environment as they showed increased locomotion (total distance) compared to controls. (B) MSEW mice show increased time spent being active (%time spent moving) during the testing period compared to controls.

MSEW alters vHIP neuronal oscillations in both familiar and novel environments

When tested in a familiar environment, no differences were observed between MSEW and control groups in theta (4–12Hz), beta (12–20Hz) or gamma (30–80 Hz) power (Fig 3D–F). Additionally, delta (1–4 Hz) power was not altered (Fig S2A). By contrast, phase amplitude coupling analysis in the familiar environment showed that MSEW male mice displayed greater modulation of gamma oscillations by theta phase (Fig 3G–H). When tested in a novel environment, MSEW mice showed significantly greater theta power (Fig 3K) as well as alpha/beta power (Fig 3L) compared to controls. Delta (Fig S2B) and gamma power (Fig 3M) were not different between groups. Phase amplitude coupling analysis showed that amplitude of gamma oscillations was also significantly modulated by the phase of theta oscillations in MSEW mice in a novel environment when compared to controls (Fig 3N–O). These data suggest vHIP theta-gamma coupling is inherently enhanced in MSEW mice compared to controls, occurring in both familiar and novel environments, but increases in theta and alpha/beta power are observed only when MSEW mice are in a novel environment.

Figure 3.

(A) Raw LFP trace from the vHIP. (B,C) High and low frequency power spectra plots from MSEW mice during 10 minutes in the familiar environment (F.E). In the F.E, there were no power differences in either theta (D), alpha/beta (E) or gamma (F) power between groups. (G) Phase amplitude coupling analysis plotting normalized gamma power as a function of theta phase in the F.E (H) Analysis of theta-gamma phase amplitude coupling in the F.E. revealed that MSEW mice had significantly more modulation compared to controls. (I, J) High and low frequency power spectra plots from MSEW mice during 10 minutes in the novel environment (N.E). (K,L) In the N.E, theta and alpha/beta frequency power was significantly increased in MSEW mice compared to controls. (M) No changes in gamma power were observed between groups in the N.E. (N) Phase amplitude coupling analysis plotting normalized gamma power as a function of theta phase in the N.E. (O) Permutation tests showed significantly greater modulation above chance for both treatment conditions (green). MSEW significantly increased theta-gamma modulation relative to controls in the N.E, suggesting enhanced theta-gamma coupling in the vHIP due to early life adversity. *p<0.05

MSEW leads to a reduction in the densities of PV+ and SST+ cells but no overall decrease in interneuron densities

Because inhibitory interneurons play roles in theta rhythm and theta-gamma coupling (23; 25), we examined their presence in hippocampus of MSEW and control mice. Analyses of the DG, CA1 and CA3 revealed that the density of PV+ interneurons in MSEW mice was significantly reduced in the ventral DG (vDG) compared to controls (Fig 4B), while the density in the CA1 and CA3 remained unaltered (Fig 4B’–B’’). PV intensity within PV+ neurons of the vDG was also reduced (Fig. S5A) suggesting that the decrease in cell density was due to reduced expression of PV, as opposed to a loss of interneurons due to cell death. SST+ interneuron density was also significantly reduced in the vDG of MSEW mice (Fig 4D), while the CA1 and CA3 regions showed no differences in this measure (Fig 4D’–D’’). It is likely that MSEW-induced decreases in densities of PV+ and SST+ interneurons are not due to cell death because there was no overall difference in the density of cells stained with the pan-interneuron marker GAD67 (Fig S5B). No significant change in density was observed in the CR+ subpopulation in either the DG, CA1 or CA3 regions (Fig 4F–F’’).

Figure 4.

(A) Image of PV+ interneuron distribution in the vDG. (B-B’’) MSEW significantly reduced PV+ interneuron densities in the vDG, but not in the CA1 or CA3. (C) SST+ interneuron distribution in the vDG. (D-D’’’) MSEW significantly reduced SST interneuron densities in the vDG, but not in the CA1 or CA3. (E) Calretinin (CR) interneuron distribution in the vDG. (F-F’) MSEW did not alter CR densities in the DG, CA1, or CA3 regions of the vHIP. (G-G’’) Images depicting c-fos activation of PV+ cells in the vDG after exposure to the EPM. (H-H’) No difference in PV+ interneuron activation in the DG or CA1 was observed. (I-I’’) Images depicting c-fos activation of SST+ cells in the vDG after anxiety-like behavior tests. (J-J’) No difference in SST+ interneuron activation in the DG or CA1 was observed. *p <0.05, **p<0.01

Mice were perfused 2h after the EPM test to allow for maximal accumulation of the protein products of c-fos, an immediate early gene used as a proxy for neuronal activation, in interneurons (44). No differences in the percentage of PV+ or SST+ cells co-labeled with c-fos in the DG or CA1 were observed (Fig 4H–H’, 4J–J’’), indicating that overt differences in the population of PV+ or SST+ interneurons activated are not likely the cause of altered neuronal oscillations and behavior.

MSEW increases the intensity of PNN labeling around PV+ interneurons

PNNs are involved in the plasticity and activity of inhibitory interneurons (45). Wisteria Floribunda Agglutinin (WFA), a plant based lectin stain commonly used to label PNNs, along with PV immunolabeling was used to visualize PNNs around PV+ cells in the brain (Fig 5A–B). SST+ and CR+ interneurons were not found to be encapsulated by WFA+ PNNs (Fig 5C–D). In the GCL of the vDG, we observed that PV⁺ cells in MSEW mice had significantly greater WFA intensity (Fig 5E) compared to controls. However, no significant differences in intensity of WFA was observed between groups in the vCA1 (Fig 5G). The overall density of WFA⁺ cells in either region was also not affected between groups (Fig 5F,H). No differences were observed in the percentage of PV⁺ cells that were WFA⁺ in the vCA1 or the vDG, between control and MSEW mice (Fig S4).

Figure 5.

(A-A’’) Images of PV+ cells and WFA+ perineuronal nets (PNNs) in the vDG of MSEW mice. (B-B’) Higher magnification images depicting high and low intensity WFA stained PNNs surrounding PV+ cells. (C-D) SST and CR interneurons are not encapsulated by WFA+ PNNs. (E) MSEW mice showed a significant increase in intensities of WFA labelled PNNs in the vDG-GCL. (F) Graph showing no differences in the density of WFA+ PNNs in the vDG-GCL. (G) Graph showing no differences between groups in the intensities of WFA labelled PNNs in the vCA1. (H) Graph showing no differences in the density of WFA+ PNNs in the vCA1. *p<0.05, ***p<0.001

MSEW increases OTX2 labeling intensity

OTX2 produced in the choroid plexus binds to PNNs in the visual cortex where it facilitates the expression of PNNs (28). MSEW mice showed a significant increase in immunofluorescence intensity of OTX2 when compared to controls in the choroid plexus where it is produced (Fig 6B). To verify if the increase in OTX2 was also present in PV+/WFA+ cells in the vDG, sections were tripled labelled for OTX2, PV and WFA and OTX2 intensity in PV+/WFA+ cells in the vDG was analyzed (Fig 6C–C’’’). Similar to the choroid plexus, MSEW mice showed increased intensity of OTX2 in PV+/WFA+ cells (Fig 6D) while PV+/WFA- cells showed no such increase (Fig 6E). Since previous studies have shown that OTX2 is not synthesized in the vHIP (46), our findings suggest that PV+/PNN+ cells in the vHIP sequester OTX2 in greater amounts after MSEW.

Figure 6.

(A) Images of the choroid plexus stained with OTX2 in the lateral ventricle. (B) Graph showing increased OTX2 labeling intensity in the choroid plexus of MSEW mice. (C-C’’’) Images showing a cell colabelled for WFA, PV and OTX2 in the vDG-GCL. (D) Graph showing increased OTX2 labeling intensity in PV+WFA+ cells in the vDG of MSEW mice. (E) PV+WFA- cells do not show increase in OTX2 labeling. *p<0.05

MSEW has no effect on the density of new neurons or proliferating cells

Because adult-generated neurons affect neuronal oscillations in the hippocampus (36), and have been linked to stress-induced anxiety-like behavior (47–49), we examined the numbers of DCX+ immature neurons and Ki67+ proliferating cells in the vDG of MSEW and control mice. No differences were detected in DCX+ or Ki67+ cell densities between MSEW and control mice (Fig S3A, S3B).

Discussion

The findings presented here confirmed previous studies that adult male mice subjected to MSEW exhibit an increase in anxiety-like behavior and activity levels (11). We further showed that MSEW increased theta power in the vHIP when in a novel environment, but not in a familiar environment, as well as increased phase amplitude coupling between theta and gamma in both settings. Analysis of cellular subtypes within the vHIP revealed MSEW resulted in reduced densities of PV+ and SST+ interneurons in the vDG, but no change in densities of CR+ interneurons, GAD67+ interneurons, DCX+ immature neurons or Ki67+ proliferating cells. Further analysis of PV+, SST+ and CR+ subtypes revealed that only PV+ interneurons were encapsulated by PNNs, which were present in increased intensities around PV+ interneurons in the vDG of MSEW mice. Reduced PV expression in PV+ cells was also observed following MSEW. These changes were accompanied by an increase in labeling intensity of the transcription factor OTX2 in the choroid plexus as well as in PV+ interneurons in the vDG. Taken together, our findings show that early life adversity produces altered neuronal oscillations in the vHIP of adult male mice and further demonstrate cellular changes in interneurons that may be causally linked to the electrophysiological and behavioral alterations.

MSEW is an adaptation of the maternal separation paradigm developed to produce reliable increases in anxiety-like behavior in adult mice (11). Early weaning by itself results in increased anxiety in both mice and rats (50;51). The combination of early weaning with maternal separation also results in increased hyperactivity in mice (11). We did not observe any changes in anxiety-like behavior in females subjected to MSEW. The lack of an increase in anxiety among female rodents subjected to early life stress is consistent with several other studies (reviewed in 52) and surprising given that women are more susceptible to anxiety and depressive disorders than men (53). These inconsistent findings raise questions about whether traditional methods of assessing anxiety in male rodents may not be adequate for females (52). It is also relevant to note that while overall significant differences in behavior were observed in 2 separate cohorts of male mice, within each study it was clear that not all mice were adversely affected by MSEW. These findings are also consistent with the work of others (reviewed in 52) and raise interesting questions about individual differences in resilience and susceptibility, the mechanisms of which will be the focus of future work.

Our data showing increased vHIP theta power in the novel environment are consistent with reports showing increased theta power and increased coherence between the mPFC and vHIP during heightened states of anxiety (20). Additionally, enhanced theta-phase-gamma-amplitude coupling in MSEW mice when in either environmental setting indicates an inherent alteration to the electrophysiological profile. Behavioral tasks can modulate phase amplitude coupling which has been implicated in sensory integration, memory processes, and attentional selection previously (54–56). Specifically, theta-phase gamma-amplitude coupling in the dorsal hippocampus has been implicated in cognitive function (57), and in the amygdala has been implicated in periods of heightened anxiety (58). Increased theta-gamma coupling has been associated with greater dysfunctional attention/arousal behaviors in ADHD children (59). While the involvement of PV+ interneurons in theta-gamma coupling has been previously established (23), our study is the first to report enhanced theta-gamma coupling in the vHIP in the context of early life adversity and anxiety/hyperactivity.

Adult neurogenesis in the hippocampus has been implicated in some anxiety- and depression-related behaviors (60; 61). Previous work showed maternal separation stress reduced adult neurogenesis in the rat hippocampus (37; 62–64), although other studies failed to show this (65). New neurons may be involved in modulating neuronal oscillations in the hippocampus as ablation of neurogenesis destabilizes network activity in vivo (36). While we did not observe differences in DCX+ and Ki67+ cell densities between control and MSEW mice, postsynaptic connections of new neurons to other cell types may be altered. Immature neurons project to interneurons in the DG and reports have documented the role of interneurons in modulating neuronal circuitry controlling hippocampal neurogenesis, as well as the activity of new neurons in vivo (38;39). It is possible that altered connections between new neurons and interneurons in the DG adversely affect neuronal networks resulting in the altered electrophysiological profile of MSEW mice.

GABAergic interneurons, especially the PV+ subtype, play an important role in the regulation of neuronal oscillations in the brain (24–27) and have also been implicated in neuropsychiatric disease (66; 67). PV+ interneurons appear to be the main generators of intrinsic theta oscillations in the hippocampus (24). In fact, disconnecting PV cell activity from the fast spiking inhibitory network in the hippocampus reduces both theta power and theta-gamma coupling (23) These reports along with our data suggest that increased rhythmic inhibitory inputs to PV+ cells may underlie the increased theta power and theta-gamma coupling observed in MSEW mice. The reduction in PV+ interneuron densities we observed in the MSEW hippocampus is likely due to decreased PV protein levels rather than loss of neurons given our data on PV expression and GAD67+ cell densities, as well as previous studies showing reductions in PV protein levels due to early life adversity in the mPFC of juvenile mice (68; 69). Along these lines, it is worth noting that experimentally induced theta bursts reduce PV expression in other brain regions (70; 71), raising the possibility that the changes we observed in PV intensity are the result of, as opposed to being responsible for, the altered electrophysiological profile observed with MSEW. Assessing this possibility will be the focus of future studies.

SST+ interneurons on the other hand have been demonstrated to indirectly contribute to the entrainment of theta oscillations in the hippocampus (25) and are responsible for preferentially driving beta (15–30 Hz) frequency oscillations in other brain regions (72). We observed an increase in alpha/beta frequency oscillation power in the vHIP of MSEW mice, which could be related to decreased SST+ interneuron intensities and possibly activity of this cell type. However, the exact role of SST+ interneurons in MSEW-induced changes in neuronal oscillations needs to be more completely investigated in future studies.

Our findings reporting increased intensities of WFA labelled PNNs around PV+ interneurons suggest altered plasticity of PV+ cells as PNNs have been known to regulate synaptic and cellular plasticity of the cells they encapsulate (45) and digestion of PNNs around PV+ cells has been shown to reduce the fast spiking properties and excitability of PV+ cells (73) In fact, the maturation of PNNs around PV+ cells in the visual cortex coincides with the closure of the critical period for the development of binocular vision; degradation of PNNs has been shown to reopen this critical period in the visual cortex (74; 75). PNN degradation also reduces PV+ interneuron activity and alters gamma and theta range neuronal oscillations in rats (29). PNN degradation has also been shown to reduce theta synchrony between the amygdala and the association visual cortex; brain regions involved in fear memory recall (30). However, no previous studies have investigated how early life adversity impacts PNNs and our findings raise the possibility that increased PNNs around PV+ cells may drive the increase in theta oscillations, as well as elevated anxiety and hyperactivity observed in MSEW mice.

Orthodenticle homeobox protein 2 (OTX2) is a transcription factor known to be important in the regulation of the critical period of plasticity in the visual cortex of developing mice (46) and its expression coincides with the maturation of PV+ cells and the closure of the critical period (35). Additionally, OTX2 has been shown to specifically accumulate in PV+ cells through exogenous transfer from the choroid plexus, the region where it is produced (35; 46). Our findings showed an increase in intensity of labeling of OTX2 in the choroid plexus suggesting an increase in OTX2 production in MSEW mice. We hypothesized that this could lead to an increase in accumulation of OTX2 in cells of the vHIP, where OTX2 is not made (46) but may work in a non-cell autonomous way as it does elsewhere in the brain (28). Our observations showed that MSEW mice displayed increases in OTX2 labeling intensity in PV+ PNN+ cells in the vHIP. Continual binding of OTX2 to PNNs has been shown to be necessary for synaptic stability in the adult visual cortex and blocking OTX2 uptake into PV+ cells in adults has been demonstrated to reduce PNN expression and reinstate juvenile levels of plasticity (28). Our results are consistent with an OTX2-mediated mechanism of action for modulating PNNs and hence, PV+ cell plasticity in the vHIP. Additional discussion of the role of cell autonomous OTX2 in stress resilience (76) can be found in the supplemental materials.

These findings are the first to report that early life adversity results in an altered electrophysiological profile in the vHIP in adulthood, which is associated with an increase in anxiety and hyperactivity. The data presented here further indicate that OTX2, PNNs and PV+ interneurons are affected by MSEW and these changes may participate in altering neuronal oscillations underlying anxiety and hyperactivity. Understanding such mechanisms is important in the identification of novel targets to develop circuit level interventions for countering the negative impact of early life adversity on neuropsychiatric disease.